 All right, so we should be live and I hope I'm audible. I've been having some major disconnects on the microphone that I'm using. But welcome everyone. Welcome. Can someone in chat see if people can hear me? Just say something. So I've been using this nice microphone, but apparently it keeps disconnecting halfway through every time. You sound a bit low. So you're saying audio up like bass. How do you mean like bass? I can see if I can change that. Like it comes with this g-hub thing. So how about this? That sounds a little bit better, but it's going into the red completely. So is this better? But I do have a very low voice, Michelle. You know me, you know me. Better, better, better, better. Worse, better, worse. I can hear myself though. Not bad. Radio voice. Yes. Yes, radio voice. We have to do that always. Anyway, welcome for joining. At least Michelle is here. Anna is here. So that's perfect. So I wanted to create some more content. It's okay. Well, we're not going for okay, Michelle. We're going for good. But like I said, it's been disconnecting the whole time. So it just, it was a pain in the ass getting everything set up in time. Perfect. Yes. That's what we're aiming for, right? All right. So welcome everyone. People are switching in. So that's really, really good. I am Danny. So I'm a bioinformatician and today is going to be kind of an AMA. I just wanted to create some content because as you guys know, I try to minimize the number of ads that are on my channel. And to do that, I need to be monetized. So I've been monetized for some time now. I think early August I reached the 4,000 watch hours. So, but if you, if you are monetized, you can set the number of commercials that you have. But if you don't stream or upload any content for like six months, YouTube will demonetize you, which means that they will turn on all of the commercials again. And you're not getting any money from that. So that's bad. But also for viewers, it's bad because it means that you have a lot of commercials. So I just wanted to create some content, do a little bit of an AMA. People can ask me questions. And if there's no questions, we'll just start programming some or I will make a nice drawing. And then that will be it. But any questions that you have about bioinformatics, let me know. So I will go to the standard layout like this. So I thought it would be nice to start off with a little drawing so that people can think of questions related to bioinformatics. And that also means I can test my drawing board. I think or I guess that the drawing board was actually the thing which is causing my microphone to be bad or not so much be bad, but to disconnect all of the time. Because as soon as I plugged in the drawing board, the Yeti was just completely gone. All right. So since there's no questions yet, I want to ask you guys, give me a topic because to draw something, I need to draw something. So at least going to go to draw mode. And I'm going to get myself a pen and I'm not going to draw with touch. So let me see if this works. So hello. All right, that works. Perfect. Cool. So today is bioinformatics. And I actually wanted to just code something, but I didn't have any inspiration. So I hope you guys as viewers have some inspiration. Like for example, I've been working on this geo data set and I would like your opinion and I can take a look. If you're having another type of data set, which is publicly available, then that's also fine. Let me know. If you have any questions in general, then that's also perfectly fine. So we're just going to do a little bit of a drawing. So in the meantime, while we wait for people to come in and people having questions. So I can tell a little bit about me. I am Danny. I am a bioinformatic. I work at the, well, not the, I work at Northumbria University now. So I'm associate professor there, bioinformatics. And my day job is coding stuff for people. It's not so much coding stuff for people. There's a lot of other things that you do as a bioinformatic. So I got a lot of questions about, so what do you do on a day-by-day basis? And kind of like, what is the life of a bioinformatic? So that is actually very, very varied. So on a day-to-day basis, I do a lot of different things. So most of my job, like I already said, is sitting behind a computer and programming. Besides that, I am also in the lab sometimes. So for example, yesterday I was in the lab. I was extracting DNA from bees. So we were trying a new protocol. Protocol didn't work. Unfortunately, we only got like five nanograms per microliter. So that was bad. So yeah, so maybe start how you get data and how you process it. Just an idea. Yeah, that's what I wanted to do. So Ibrahim Musa, welcome. Welcome to the stream. Yeah, so on a day-to-day basis, like I say, I program, but I also do generate my own data. So the university gave me a little bit of startup funding. So one of the things that I'm currently researching is to figure out if we can reconstruct the genotype of the queen from sequencing individual bees. So I was in the lab yesterday extracting DNA from bees. So that's just buying a kit and following the protocol. Unfortunately, it didn't work because we messed up the protocol at the beginning. But that's generally how you get your data. So your data is going to the lab, either yourself or a biologist goes to the lab. They follow a certain protocol to either extract DNA or extract RNA, extract lipids or metabolites from their samples. And then they do some kind of hype throughput technology. So let me kind of make a little schematic drawing for you guys. So we start off with doing something in the lab, right? So we have like an Eppendorf tube with a little bit of liquid in there. And then what happens is that we put something in there, right? So for example, I'm currently working on bees. So let me draw a little bit of a bee. Looks like this bees are yellow. So let me make it yellow as well. So what we do is we take the bee and then we crush it. So we put it in a little mortar and pestle, right? So we take the bee, we put it in the mortar and pestle and then we crush up the bee. How much time do you spend in dry lab versus wet lab? So that really depends on the week. So yesterday being in the lab was or not yesterday, but the day before yesterday or last Friday was the first time in like six months that I was in the lab. So I generally do around, I would say like two to three days of lab work every half year. And that is because I'm more or less a computational biologist. I have a master in molecular biology. So I know all of the lab techniques and I can do all of these kinds of things. But for me, the part of lab work is really, really small. And that's just because I only go into the lab to gather either my own data or to help other people out. But I know of other bioinformaticians who spent up to like 50-50. So they will have like two and a half days of working in the lab, two and a half days of sitting behind the computer and analyzing the data. It might also help to give your opinion on the difference between bioinformatics with computer science background and those with a biological and life sciences background. Yes. So I have a computer background, right? So my background is computer science. I did that at university and I absolutely hated computer science. So I quit doing computer science and did biology. So I just have a bachelor in biology, master in molecular biology. And then I went back to become a bioinformatician. So during my PhD, I did a lot of programming and combined with the biology work that I do. So I don't think that in the end, there's a lot of difference from someone who has a biology background or from someone who has a computer science background. It makes a big difference when you start off. Because of course, if you do a master or if you do a PhD, imagine you do a PhD program, you come in with computer science background, then you don't have the biology background. At that point, you will spend more time in the lab getting to know all of the lab techniques, getting familiar with being in a lab. But of course, that is because you already know the programming part. On the other hand, if you're coming in with a biology degree, then of course, the programming is going to take up much of your time. But of course, after working for four years on a PhD and then working in bioinformatics itself, either at a university or at a company, the difference becomes less and less because the background that you have should kind of merge into a general background where you have 50% biology background, 50% programming background. It really depends also on what the company wants you to do. Right if you're working for more or less a software company designing software for bioinformaticians, then of course, you're going to spend 90% of your time, 99% of your time programming. But if you're working in a lab where you do high throughput sequencing or where you do metabolomics or lipidomics, then of course the lab component is going to be much, much bigger. But Ibrahim, I don't think that there's a difference in bioinformatics depending on where you come from. I think there's a difference between being a data analyst and being a bioinformatician. Because a data analyst generally does not go into the lab while a bioinformatician generally goes into the lab sometimes, not always, but there's always a lab component to it or component working directly with data. Hello, I have some software developer or data analyst wants to get into bioinformatics. What services can he provide as a freelancer for bioinformaticians? Gig on Fiverr, which one would be most valuable? So if you are a software developer and you want to get into bioinformatics, you need to start learning the biology because you already have the software background. So you will probably have to learn the biology part. There's always freelance. There's a lot of people in bioinformatics who work in a lab or who work in a bioinformatics environment and they write very basic analysis software. So they write analysis software for their own kind of analysis and they end up with academic type level software. So there as a software developer you can really contribute a lot by professionalizing the software. But I think that if you want to get into bioinformatics, the main thing is learn the biology because it's not for nothing bioinformatics. The bio is first and then the informatics is there to support it and then start looking at what are some open questions in bioinformatics. So there's a lot of things that we haven't solved yet but bioinformatics is really a huge field. So figure out what you enjoy if you like doing genetics or if you like doing data analysis. If you for example like writing software there is also a place for you in bioinformatics. So I think that there's enough open questions that anyone who has a specific interest can find something that interests him. But I would say being a bioinformatician is being very flexible. So on a day-to-day basis like I say some days I'm in the lab, other days I'm programming. I of course work for a university so I'm writing money. So I'm writing just project proposals, submitting them. I'm writing papers. I'm reviewing papers. I do teaching and I support PhD students. So those are more or less all of the different components that are there to my work. And of course talking with people as well. Alright, so no more questions for now. So I will just go back to kind of what I did in the lab and what we do so just to finish my drawing. Right, so we have our B's over here. I have my mortar and pestle. So the B goes in, we crush it. Thanks so much for doing this. Yeah, no problem, no problem. So I take the B, I crush it and we put it in this little cup. Right, so the whole crushed B goes in there and then we just run a protocol to extract the DNA. Right, so this is a kind of two-hour protocol where we get DNA and this DNA is in solution. I'm very bad at drawing DNA, but now we get DNA. So this failed yesterday because we tried, we didn't get enough DNA, but if you get enough DNA and you get good concentration of DNA, then you can go on to the next step. So, alright, question in chat. Hello doctor, what do you think are the skills needed for someone who wants to follow a path in pathogen genomics? So skills needed for pathogen genomics. I would say you probably have to have some lab skills then. So you should be able to work in a moon suit or I don't know how people actually called it, but you should be comfortable wearing PPE for like six to eight hours per day because if you work in pathogen genomics you are going to work with pathogen, so you have to extract those. Yeah, call it a hazmat suit, but it's generally like personal protective equipment, so it doesn't have to be a full hazmat suit, but you have to be comfortable working in a lab wearing PPE. Besides that, of course, you need to be able to understand the data that you are gathering. So you will have to learn something like R or Python to program little scripts to do your own analysis. And besides that, you have to have a relatively good understanding of genomics. Pathogen genomics, so you have to know what DNA is, you have to know what RNA is, and you have to kind of be able to interpret the changes in the pathogen that you see on a genomic level to how the pathogen behaves. Like, hey, if you see mutations, you have to be able to figure out, okay, so I have pathogens from different patients, some patients become very ill, other patients are very infectious, and then you have to be able to couple the phenotypes that you see of the pathogen in the patients to the pathogen genetics, so to the different genes. So there's a modeling component to that as well, being able to kind of model the proteins of the pathogen and then making predictions on how changes in the DNA will affect protein structure. Do you have any advice on landing that first-job partnership, volunteer-ship position for all the pharmacology grad currently studying MSC and bioinformatics? So I always say, look for a professor that does some work that interests you, and just do kind of cold calling, so just write to people that you think are doing interesting work. And when I say write to people, don't send them a standard letter, right? Send them a really nice email where you say, oh, hey, I read your paper. I find that this is really, really interesting, and tell them why you found the paper interesting. Tell them what interests you, and then in the end close with saying that I would love to have an opportunity to work with you, and that is one of the ways of getting it. Of course, besides that, you have to have a certain portfolio that you can fall back on, right? So you have to have a GitHub repository somewhere where you can show, oh, I did all of these things. I can analyze data coming from next-generation sequencing, or I can do protein modeling, but that is kind of how you would be able to get a first job, a first internship. So kind of build up your portfolio. Generally, you do that during your masters, right? Because in your masters, you do one or two master projects. So these master projects make sure that the code that you produce is code which you are really, really proud of and that you can show off. But for the rest, I would say, science is very driven by PIs, by people. So read papers, find what interests you, and then definitely just contact people. But when you contact people, make sure that you don't contact them in a standard way. If I just get a standard email saying, well, hey, Dr. Denny, see my CV attached. I hope you have a PhD opportunity or a volunteer or that's not going to work. For me, what works is say, I saw your YouTube channel, I like it. I read these papers that you, and I have some questions about these papers, or I think that you could get more out of the data, right? So hook them, hook the PI, and then they will probably be interested in working with you. Hi, what current BIOinformatics research do you think is the most interesting in your opinion? So in my opinion, the most interesting that I'm currently working on is the longevity research. I think that longevity research is getting to a point where it starts becoming real. In the past, people did a lot of work on longevity, but they generally had very small sample sizes, right? So they would take like 50 mice or 200 mice or 400 mice. But since sequencing and expression data and metabolomics and proteomics is starting to become so cheap, it is now possible to do kind of large scale, real research into longevity in animals which are relatively close to us, right? There's a lot of good work being done in Seeligans and that has been going on for the last like 40 years. But of course, the translational research from Seeligans to humans is really, really difficult. So doing aging research in mice allows you to really kind of get candidate genes for aging that are useful and that can be translated into humans. And of course, doing mice also allows you to do drug interventions. Look at what drug interventions are there. But this is just personal. So I think that that is one of the kind of new fields of bioinformatics where there's still a lot to gain. I wouldn't like to just work on genome-wide associations in humans that doesn't really do it for me. I really want to make a difference. So and I think there I can still make a difference like contributing to genetics, gene-by-e interaction related to aging and diseases that occur at like an old age. So things like Alzheimer's, Parkinson's, all of these diseases have been a mystery for a long time and we can now start to interrogate these diseases because data becomes so cheap to generate. Hi, are you working with mass spectrometry-based proteomics data? If yes, can genomic tools or packages be used for proteomics in your opinion? So yes, I have a good friend in mass spectrometry who kind of runs into my office three times a week. He is not so much interested in proteins. He does a lot of lipidomics and metabolomics work. But I think that every level of the genome or of the omics levels, right? So if you start with genomics and epigenetics, transcriptomics, proteomics, metabolomics, lipidomics, all of these levels, there is a lot of similarity between the data that you get. But you can't just take tools from one level and translate them into another level because from a statistical point of view, distributions are different, right? If you are measuring RNA molecules, you're generally, especially if you do sequencing, you're generally creating count data when you're doing RNA-seq. However, if you do proteomics, you generate kind of relative abundances, right? Because you do not count the number of exact proteins that are there. You just count how many peptides are in a mixture and then you're assigning these peptides to proteins. So there is a big difference in having absolute abundance data, which is more or less what you get from RNA-seq data versus when you do proteomics where you get more or less relative abundances and that those things have their own type of workflow, own type of pre-processing. So I think that that's one of the things also what makes bioinformatics really interesting because bioinformatics is a field where you are working with people that do all kinds of different analysis, right? So you're working with people who do epigenetics, you're working with people who do mass spec, you're working with people who do proteins. And all of these levels have a very similar question because people generally just want to know which ones are up-regulated in a disease state and which ones are down-regulated in healthy or the other way around, right? So it's always kind of a contrast. But the data that people have is very variable. So it really depends on if you have absolute data, if data counts, if data follows a normal distribution. So I don't think that you can very easily take genomics tools and apply them to proteomics. But I do think that the approaches that people take in genomics are translatable to proteomics, right? You will always have first kind of a QC step, then you will have a data transformation step, and then you will have a kind of statistical analysis step, and then in the end you have a result step. But yeah, I think that in that sense that you can't just take a package which works for gene expression and then put proteins in there. What is your personal favorite thing to do at work and current research interests? What is my favorite thing to do at work? Well, I love lunchtime. It might sound weird, but lunchtime gives you an opportunity to just sit down with people and talk about what they are doing. And I always get a lot of new ideas from just talking to people, seeing what they are stuck with in their research or what they're talking about. But I think you mean it more like what are my favorite things to do at work. So at work, I like programming. The issue is when you are getting out of your postdoc and getting into kind of assistant professor, associate professor or even professor level, a lot more of your time is spent on things which are not your favorite. You have to do a lot of administration. You have to do a lot of marking or grading of papers, which is important, but it's not my favorite thing. So my favorite thing is to just sit down, put on some headphones, put on some music and just be able to program for a couple of hours. Besides that, I love teaching as well. So every time that I can be in a classroom talking to students, that is one of my favorite things to do. Current research interests. Yes, so my current research interests are twofold. I am working on a very interesting aging project where hopefully very soon after three and a half years of work, there will be a paper on BioArchive. And besides that, I love doing genetics. So I love genetic puzzles. So figuring out stuff which we don't know yet in genetics. Are you looking into longevity stuff and have you seen that man who is trying to reverse his biological age? I haven't seen the guy that's trying to reverse his biological age, but I think that everyone kind of wants that. So yeah, aging is really hip right now, but there's a lot of snake oil being sold as well. Although there are some good anti-aging drugs coming along. So it is getting to be a real field instead of just a field with just snake oil there. Hi, Danny. What potential of drug discovery do you see in the future as a prospective field? Yeah, so drug discovery is not really my thing. I work in genetics, but I do think that drug discovery is very important. Unfortunately, I don't know a lot about drug discovery, but I do think that drug discovery is going to continue to be a field which is actively researched. I see a big opportunity for using existing drugs for new fields. So kind of the recycling of known drugs to other fields. And I'm very interested in seeing how things like psychedelics are going to work on treating anxiety syndrome and all of these things, because I think that the connection between mental disease and drugs is something that is kind of barren. There's still so much research to be done there. Because we know relatively little about the brain, but I think that that is one of these fields of research where there's still a lot of potential for new drugs being discovered and a lot of good that we can still do. Hi, Danny. I'm a great admirer of your videos. All right, thanks. I'm happy. It's always good to hear people enjoying the videos. The videos that are on my channel are kind of a culmination of like eight years of teaching. So when I started in Germany, I started with kind of completely rebuilding the courses that were there. So I gave two courses, Bioinformatics and Art Programming. And so I started in Germany in 2014-ish. So the pandemic came around in 2020. So after like six to eight years of giving the courses live and seeing what works and what doesn't work, I was actually able to or I was forced to do them online due to COVID. And I'm really happy that I recorded everything and had the opportunity or that the university allowed me to put them on my YouTube so that people can enjoy them. And it's still one of my proudest achievements, I think, getting all of these videos out and seeing the reaction like I never expected to get like 4,000 plus followers from just posting Bioinformatics videos. And I'm still flabbergasted about that. So I thank you guys most for liking the videos and commenting and just generally watching them. How was your experience working in Germany? I have a very, very positive experience working in Germany. I was part of a great group. We did very good research and yeah, I liked working in Germany. Berlin is a great city to live in, although Berlin is not really Germany-Germany. But I have a very good experience working in Germany. One of the nice things about working in Germany is that you actually get to learn a second language because academia in like a lot of places is very English driven. So people talk English all of the time. But in Germany, that's not really true. In Germany, there's still a lot of German being used. So you get to learn an additional language for free. One of the things that I don't really like about Germany is the way that they treat people in academia with their kind of temporary contract system and the fact that there are no real permanent positions in academics for people under professor. And I think that that's really bad. I think that there should be more tenure tracks. So the opportunity to work for five years show that you are a good researcher that you can teach and then be eligible for a permanent position. But that's not how the German system works. What part of your daily work would you like to outsource to an R or Python programmer? In my opinion or from my perspective, I would not. A question earlier I mentioned that I'm actually really happy if I can program. So the programming part of the job is getting less and less because I do more admin, I do more teaching, I do more writing of funding. So if I get an afternoon of not having to do students or not having to do admin or not having to do funding, I really, really enjoy programming in R and Python myself. But what part of your daily work would you like to outsource to an R or Python programmer? I don't know. I find this a really difficult question because I think a lot of people that are in bioinformatics started doing that because they actually really, really enjoyed doing bioinformatics and doing the programming, sitting behind the computer solving questions. So I think that not a lot of people would be willing to outsource a lot of R and Python. Although I am outsourcing a lot now because when I have an idea, I just don't have the time to implement it. So the ideas go to undergraduates or graduate students or PhD students to work on. So I'm more of an idea generator at the moment than that I get to actually implement those ideas myself. Hi, Danny, your series on RNA-SEQ was the best in YouTube. Please do upload a video of exome sequencing data analysis. I'm looking for collaborations as well as a postdoc. Please do guide me in this regard. All right, I will write it down. I don't think that there's a big difference between standard RNA-seq and exome sequencing. All of the sequencing techniques, be it DNA sequencing, they all follow a very, very similar pipeline. And of course, the RNA sequencing data from scratch is an interesting way of looking at it because in a company, no one would write their own pipeline from scratch. They would just take a standard pipeline and they would just push their data through and they would spend all of their time looking at the analysis part. But I thought it was really interesting to kind of do this so to show people how to do every step by yourself. Downloading the data, adapter trimming, going through all of these steps that you have. And I think it worked out really well. I'm really happy about the RNA-SEQ series. Although I did get some pushback on Reddit where people said like, ah, you shouldn't do it like this. You should teach people to just use things like DSEQ2. And in a way, I do agree. If you just want to make production and analyze your RNA-SEQ data, then yes, you should just take an off-the-shelf analysis pipeline like DSEQ2 or Seurat or whatever you want. But if you are a researcher and you want to know how these things work, then I think it's a really good series for that. So I'm very happy that you enjoyed the series. Hi, Danny. If you had chosen, would you prefer to work in a team of mostly biologists or a team of mostly bioinformaticians and computational biologists? That is a difficult question. Because during my postdoc, I was mostly the only bioinformatician in a group of biologists and currently I am setting up my own group. But I work a lot together with other bioinformaticians now a lot more than I used to. And before I had my postdoc during my PhD, I was in a pure bioinformatics lab. So I have both experiences and I don't know which one of the two I like more. It's a difficult question in that sense. If I had to pick one of the two, I would say it's probably more fun to work with purely biologists and be like one of the few bioinformaticians that supports like 10 or 12 other people doing their research. And this is because you learn a lot more from it. It forces you to be very efficient with your time. It forces you to be very efficient with kind of understanding other people, understanding biologists and translating that into code requirements or in analysis requirements. Being in a group of pure bioinformaticians, however, has the advantage that you develop your skills much more. So you become a much better programmer. So instead of someone who is writing scripts all day, analyzing very diverse sources of data, being in a group with pure bioinformaticians allows you to kind of get your skills to a higher level. So if you really want to improve your R coding or improve your Python, then it's really, really nice to have other bioinformaticians surrounding you. How's your grind on dwarf fortress going? Actually, really, really good. I'm after spending 300 hours plus in dwarf fortress over the last weeks and months. I'm actually starting to become really good at dwarf fortress. I started yesterday in an evil biome where, like, if something dies, it reanimates. And I'm up to 92 dwarfs and none of them have died yet. So I'm actually getting better and better at keeping my dwarfs alive. But I'm sorry for you guys because it does mean that I stream less. But I think, yeah, it's a fun game, right? I've been playing dwarf fortress for, I would say, 15 years or something. And it's good to see the atoms now making a kind of cash grab. That sounds really negative, but I actually mean it very positive. I'm really happy that the developers now put it on steam, got a boatload of money, and I'm really happy about how it's going. And I like the new graphics update as well. Are you going to play Diablo 4? No, unfortunately, I will not do that. My gaming is very limited to strategy games and Project Zomboid. So the games that I play the most is like Factorio, Dwarf Fortress, Rimworld, Project Zomboid. Would you like to stream your gameplay sometime? I actually do. I actually stream on Twitch. So on my Twitch channel, you can see me brutally murdering dwarfs or getting my colony blown up in Rimworld. So I do stream, but I try to keep the YouTube channel for purely bioinformatics programming things. And the Twitch is more for I want to play some games and talk about either bioinformatics or Dwarf Fortress or these kinds of things. So, all right, a lot of questions. Like I thank you guys for actually showing up. I was a little bit worried that I would be just sitting here for like five minutes just staring at the screen. But I'm surprised by the amount of questions and the quality of that. I'll check out the Twitch channel then. Yeah, I think you're just too late. I think my latest Dwarf Fortress run because Twitch only saves the video on demands for like two weeks. So I think you just missed it. But I might play a little bit on Twitch tomorrow since tomorrow is a free day for me since it's a bank holiday here in the UK. So and I'm getting confident enough to stream myself getting on an evil bio and not getting slaughtered within the first like five to 10 minutes. So thanks for asking. Cool, cool, cool. All right. So since there's no further questions at the moment, I will continue to kind of explain kind of what I'm trying to do with my my B project. Oh, question comes in. So do you sometimes feel the terminology about bioinformatics is a bit confusing specifically around go enrichment using over representation analysis and GCEA gene center enrichment analysis like a rank system. Yes, bioinformatics is one of these fields which falls into the classic trap of reinventing or redefining words in their own meaning. And I think that that's one of the things that you can it's the it's the thing that you can see if someone is a really good bioinformatic because if people are able to translate different words between different domains and still be understandable to the people from that domain, then I think that that is the hallmark of a good bioinformatic because a lot of words are actually used differently from a biology perspective relative to a computational perspective. So a bioinformatician is really someone who sits between biology on the one side and between it and HPC on the other side. So I see a bioinformatician as kind of an intermediate. You do a lot of translation work. So you're telling biologists that you can't do that because this is not computable, right? But bioinformatics does fall into the trap of reinventing the wheel and reassigning terminology in a way. And I think we should be very, very careful about that. But bioinformatics is really confusing sometimes and especially in the AI field at the moment. So there's a lot of AI coming into bioinformatics which brings in another dialect, right? You have the standard kind of programming dialect. You have the biology dialect and you have the dialect of the people who are really computational. So doing IT and maintaining clusters and HPC and cloud. And at the moment you see a lot of AI terminology come in as well. And I think that we as bioinformaticians need to be really, really careful. And when we write papers, we need to be very, very good at writing down what we did and how we did it so that research continues to be reproducible. How would you compare living in Germany and the UK? It's different. The UK is very different than Germany. Like, I never knew what a meal deal was. But now I know that in the UK, everything's a meal deal. So if you buy a sandwich, then the sandwich is way too expensive. And that is because they just assume you're also going to take a Coke and you're going to take some chips with it. So the combination of the three is normal. But the sandwich itself is way too expensive because of the fact that everything's a meal deal. Those things did not exist in Germany. But comparing living in Germany and the UK, it's difficult, right? I lived in the center of Berlin. I now live on the outskirts of Newcastle. I think the cat has a very clear opinion that the UK is a lot better. And that is because Oscar now has an outside, so he has his own garden and he can play with the other neighboring cats. While in Berlin, he was limited to being in house and being on the balcony. But there are a lot of differences. But in the end, it's very much similar, right? It doesn't really matter where you are across the world. It's all Western civilization, right? So supermarkets in Germany are very similar to supermarkets in the UK. Do you keep the bees yourself? Amazing little creatures. No, the bees are actually not mine. So I'm fortunately enough that some of my collaborators here in the UK, they have contact with beekeepers all over the north of the UK and Scotland. So they go out and they collect bees from the hives of beekeepers. So we have currently around 600 bees in the freezer. The 600 bees come from 25 colonies and then two time periods. So every, well, I would say every half year, bees are sampled from the new hives and these bees are brought to the university. They are part of a big microbiome project to look at gut microbiome of bees, but I'm more interested in the genetic puzzle. So taking the individual drones and seeing if you can reconstruct the queen's genotype out of those. Hi, Danny. Hi, Chad. Hi, Solomon. What do you suggest as bioinformatician to learn R or Python first or a combination of both? I would not learn a combination of both. I always advise people to learn R first. And that is because R is the easier language in my opinion. Because it just has so much built into the language. So R understands statistics and there's a lot of useful data wrangling tools that you can do with R. So learn R first, become proficient. So proficient means that if you have a question on a Saturday afternoon, you can start coding and you can actually answer that question yourself. So, hey, if you have an hypothesis, you think like, oh, I want to know if people on Facebook, if their friends are kind of clustered together, then you can actually figure that out by coding, by downloading data from Facebook, by spidering it, doing statistical analysis. But it's difficult. But I would not ever advise anyone to learn two languages at the same time. Because you start getting confused, you start mixing them up, and in the end, you'll take more time becoming proficient in both of them. So pick one of the two, either R or Python. Learn it till you are proficient and then learn the second language. It will make your life a lot easier. Hi, Denny. Thank you for your help with the questions I posted on your last video. No worries. I always try, if someone comments on a video or asks a question on Reddit, I always try to take some time to help people out, because people take the time to write up the question and they take the time to watch my videos. So I feel kind of obligated to help you guys out. I was able to figure out my script and it runs smoothly without stopping. Okay, perfect, perfect. So that's good to hear. You were the one that stayed up until 6 in the morning trying to debug a script. Mad respect for that. You shouldn't do that. If stuff doesn't work, spend a couple of hours on it, but don't lose sleep over it. So don't do that. Just go to bed at 12 and start fresh in the morning. Generally, that works for me the best. Hi, Denny. Thanks for taking the time. Do you see an overlap between healthcare informatics and bioinformatics? Also, have you seen physicians undergoing bioinformatics training? Yeah, so healthcare informatics is different from standard bioinformatics or what we consider bioinformatics. So it's more writing apps to support healthcare. But like I said, when I was doing my PhD in Groningen, part of our group, so our group was the Groninger Bioinformatics Center and the Groninger Bioinformatics Center actually is shared between the faculty of biology and the hospital, so the UMCG. So I don't see this big divide between the two. Of course, there are different skills and different tools which people use in healthcare informatics, which is mostly app design and patient data. But there is a big overlap between healthcare informatics and bioinformatics. And as such, I would say that yes, I've seen a lot of people that transition from being a physician and learning computation or learning programming on the side, generally not to become bioinformaticians themselves, but generally they do this to understand or be able to talk better to existing bioinformaticians inside of their hospital. So they learn how programming works to be able to talk to programmers that are already within the healthcare system within their hospital. I'm a microbiologist and I do not have much bioinformatics or programming knowledge, but I newly started to learn R and Python like Amandico. I wanted to study and learn about biofilm, forming genes and some bacteria that I'm interested in. But there isn't much knowledge or on the papers about it. I want to find the genes related with formation and they also want to design a specific PCR primer for that gene. Where should I start? What is the path that I should follow? I would say first make an inventory of biofilm formation. As far as I'm aware, there is a lot of research on biofilm formation, especially in a transplant setting. If you get a new hip, then the main problem there is that if the hip that goes in is not clean enough, biofilms will be there. I would look to see what kind of data sets are out there. For me, as a bioinformatician, I always like to start with a data set and then start replicating what other people did. By replicating what they do, you start seeing where things can be done better and where there might be things that you would have done differently. I would Google a lot, learn about biofilm forming bacteria and which genes are involved. See also who the people are that are studying this field. Try and find someone who is publishing about biofilms in high-impact journals. Send them an email, like I say. Be friendly, say that I loved your paper, I loved the analysis that you did. However, I did not understand this, this and this, or I wanted to replicate your work, but I can't find the data set because it's not available or I'm having issues with it. Start there. Try finding a mentor that can help you progress. If you want to design specific PCR primers to that gene. PCR primer design is a field of its own. I have some videos about PCR primer design. They're in the bioinformatics lecture course. I can't remember exactly which one it is, but PCR primer design is not too hard. You can pick that up in an afternoon. Just watch the videos, do the assignments that are there because in theory it's just finding a stretch of DNA which is unique to your target. So have which is unique to the bacteria that you're targeting and then just target a 20 base pair primer on the left side of your target, on the right side of your target. And when you're designing primers to measure gene expression, you generally make axon spanning primers. Although for bacteria that probably won't work because bacteria of course don't have an internal axon structure. Alright, John Cardeno, R or Python? So for me it's R by far. I've been dabbling in Python a little bit last week because I was trying to do some little website with Flask and stuff. And I actually wanted to learn more about large language models. And of course for that you need Python because those are all written in Python. So R for sure. T or coffee? Coffee without a doubt. The people in the UK will always call me crazy but I am a coffee junkie. So I have my first cup of coffee as soon as I wake up and then the second one comes afterwards and then I go to work and then I have another coffee. And then at 11 I probably have another coffee all the way up until like 3, 4 in the afternoon. But I don't drink coffee in the evening. In the evening I prefer tea. Pineapple on pizza? No. No pineapple on pizza. I see my moderator also said that already. So no. It is a crime, a crime to do pineapple on pizza. Can we get a viewer reward for this stream? Unfortunately not. But I was starting to do a drawing so it's kind of a viewer reward for all of you. But no, for the viewer rewards you have to be on Twitch. So you have to sit through me killing dwarfs to get a viewer reward. Do you know any labs, bioinformaticians working in the field of sports performance, sports science? Yes. Here at Northumbria we actually have a big sports department right across the road from us. And definitely look at Northumbria University. There's a sports department and they have bioinformatics as well. So again, this is a slightly different type of bioinformatics than bioinformatics for biology because here it is mostly about athletes. So they look to see how people can perform better than they normally perform. So they do a lot with bioinformatics on data that you get from wearable gadgets like the iWatch, the Fitbits and these kinds of things. So they do a lot of tracking of like heart rate, oxygen saturation. But they also do a lot of research on like lactic acid in muscles. And they of course, they work together with the Newcastle United Sports Team. So it's an interesting collaboration there and they do bioinformatics to analyze their data as well. Kofi is the true lifeblood of academia. Yes, yes, yes, yes. And no, Misha, we're not going to do a puffer fish again. There is already a puffer fish that has been done. Kutampasi, bad timing for me, have a bioinformatics exam tomorrow. I should be studying for the exam, but the stream is fun. Hopefully I'll be able to join the next one. Yeah. Now, I actually thought about making this kind of a regular, irregular thing where once every, like once a month, perhaps once every two months to do a stream like this, just to kind of fill up the content that I normally have because the three hour streams where I teach you guys something new, those are really, really fun to do, but those take a lot of time to prepare. So it is a nice break to just sit here on a Sunday and talk to you guys and give my opinion on stuff. John Kudena, super interesting thanks. I checked him out. Yes, and there are more sports bioinformatics things as well, but it's slightly away from my core, right? So my core is really genetics. Mice, animals have production things. So I'm not too knowledgeable about sports, but I know that Northumbria has sports and other universities probably have that as well. And if you have a sports department which works with a bioinformatics department, then generally they do good or interesting bioinformatics research, but it's more wearable that it is real like genetics and sequencing. All right, I will take a very short break. I will be back in like four minutes. I just want to get something to drink and do a quick cigarette. So keep posting questions in chat. I will get to your question. And if not, then I will actually just continue talking about my B project a little bit. Along with the long videos, probably concise review video would be great if you can make them at least an idea for future videos. Yes, I actually thought about scaling down a little bit to like 20-minute videos and check your latency webcam microphone. All looks good actually on my side, Misha. What's your issue? Are you getting like stutters? Or does the longevity aging research focus on telomeres more? No, we are actually... So I can talk a little bit about the aging project instead of the B project. So the aging project is focused on mice. So we are using UMHAT3 mice. So they are mice from four different founder mice. Your audio is a little bit... Oh, okay, okay. I can see if I can fix that during the break. But about the longevity aging research. So the longevity aging research that we do is just on mice and we find genetic determinants of aging. And these genetic determinants of aging are more or less, it's just basically QTL slash association analysis on a very large, almost 7,000-mouse population. So this aging research that we're doing is there to kind of find which genes or which regions of the genome are associated with aging and what can we learn from that. So we don't focus on anything specifically. We are taking a data-driven approach. So we are not coming in with a hypothesis, but we are more or less generating data and then seeing what the data tells us. We are hoping to see things like telomeres, mitochondria, and these kinds of things to put it in a better... or to put it in context. But we'll have to see what comes out. So it's a data-driven approach. So we just have the largest population of aged mice in the world. And then we do analysis to figure out where in the genome there are determinants for how quickly they age and if those determinants are male or female specific. So it's going to be a couple of more weeks, but I'm hoping that in like a month, month and a half, we will probably have a bio-archive version online. And then we're going to see which journals are interested in publishing it. All right, I will get something to drink very quickly, do a cigarette, and I will be back in five minutes. So at 2.15. What does it mean AMA? AMA means ask me anything. So you can ask me if I like pineapple on my pizza. You can ask me about bioinformatics. So that's kind of the idea. Here is just you guys are free to ask me anything and I will do my best to kind of answer your questions. All right, I will be back in five minutes. So bear with me and I will see you guys soon. All right, so I'm back. I've got myself some water so I can continue asking or answering questions. So AMA, yeah, ask me anything that whatever you want, if I know the answer, I will tell you the answer. If I don't know the answer, I will not make up something, but I will tell you I don't know. But you can get my opinion on anything that you want. So I'm actually really surprised by you guys. Let me start off by saying that I'm really grateful for all of the questions that I've gotten so far. I was a little bit worried that no questions would come in and I would just sit here talking to myself. But how did you get started in bioinformatics? What first interested you? That is a very good question. So when I was studying during my master of, so my master is in biochemistry, is in molecular biology, I got into the lab of Oscar Kuipers. And Oscar Kuipers does interesting research on bacteria. So they do bacteria for cheese production. So they work on all kinds of bacteria that are used in cheese production to kind of figure out which bacteria gives which properties to cheese and how does different bacteria impact fermentation rate and these kinds of things. So they had an open question on how to predict operons from sequence data. So they had a lot of data on which genes are co-expressed. So on a single messenger RNA or an mRNA, right? So I used random forest as a machine learning approach there to build a classifier to see if two genes are transcribed together or if they are transcribed on a single operon or not. So after that I wanted originally to follow not the research master but to follow the kind of business part of the master in Groningen because the master in Groningen is split into two. You can do kind of a research track where you do two research projects or you can follow the business track where you do one research project and then you write a kind of an advice report. So you go more into the business side of things. So I originally chose the business side of things but after doing this first project with Oscar I really enjoyed using my programming skills for something that I found interesting because I had studied computer science before. I didn't really like that because it's mostly about efficiency and calculating prime numbers, building websites, these kinds of things and that doesn't really interest me. So the thing that got me into bioinformatics was the realization that I could use my programming skills to answer interesting questions, questions that biologists had and that biologists couldn't answer because they didn't have the ability to do either the programming themselves or were not able to use things like clusters or cloud computing and since I had a background in that I found that really exciting. So my second project I did with Richard Johnson and that really changed my life. So Richard is the professor that I probably owe the most to because due to him I got really interested in bioinformatics and genetics and that is kind of where I got hooked. So I really got hooked by the things like mouse genetics and these puzzles on figuring out how pieces of DNA are inherited in family structures and how they control phenotypic variation that we see. So that's kind of when I got hooked in bioinformatics and started more or less become a real independent researcher. How to make money in bioinformatics? Several different ways you can make money. So I'm an associate professor, right? So I get paid by a university to do education. So I give lectures, I do research, I write research projects but I just get a fixed salary. So I make my money from that. I'm not planning on becoming rich because in academia you don't become rich. But there's a lot of different ways that you can make money in bioinformatics. You could develop a laboratory information management system. So a system where you kind of sell your software on tablets that people use in the lab to keep track of the experiments that they do or to keep track of the animals. Nowadays a lot of companies are kind of building these automated pipelines for research. So I'm thinking about companies like Galaxy and Ledge.Bio, these kinds of companies they make money by kind of being an intermediate between biologists doing research and high throughput infrastructure that companies like Amazon or Microsoft or Google have. So they kind of sit in between biology or academic biology on the one side or healthcare or hospitals and on the other side the big cloud compute providers. So that is where a lot of people in bioinformatics make their money. A lot of people are like me and they're just working a day job. So they're making money due to their ability to teach and doing research on the side. If I want to study a new undergraduate it would be bioinformatics. That's very good to hear. I think bioinformatics is a very interesting field and the nice thing about bioinformatics is that you can get into bioinformatics with very different undergraduates or even masters. So I've seen people from physics, mathematics, biology even at a later age re-schooling themselves and going into bioinformatics. So I think that that is perfectly possible. And you don't have to start out by doing bioinformatics on a university. You can even after you finished your master and have been working for a couple of years retrain yourself. So learn a programming language and kind of get into bioinformatics. Or of course if you're already a programmer working for Amazon or Google or whatever you can retrain yourself and you can learn the biology and you can get into bioinformatics that way. Advice on picking a thesis dissertation topic for distance learners, no lab, own generated data set. I think that that's really hard. I think that you're better off having at least some guidance from a professor or someone that will help you pick a good topic. Help you explore that topic but also to teach you the things that you need to learn to write a good thesis. Like scientific writing is not like normal writing. So it is really something that you need to practice. You need guidance on. And like I said without Oscar Kuipers, without Richard Johnson, I wouldn't have gone so far in my career and had along the way I've had many, many different teachers like people that I still work with on a day-to-day basis. So I think it's very hard to really excel as a distance learner. But there are possibilities of course, right? If you are really good at programming and you're really good at writing, then there's so much free data available that you could make a thesis out of publicly available data. But you have to come with your own idea and say like, okay, so how we have all of this data available on RNA sequencing which is available in the SRA and I have this hypothesis and I take all of that data and I try to validate or invalidate the hypothesis that I have. So I'm not saying it is impossible, but I'm saying that it will be much easier when you have someone who is known in the field or someone who kind of knows what the current issues are to guide you through. I, biochemistry students, so how can I connect bioinformatics with that? So biochemistry and bioinformatics are not too far apart because chemistry is of course not biology, but biochemistry is very much biology. So if you think about biochemistry, right? You think about how to use proteins and enzymes to produce things or to recycle waste or these kinds of things. So bioinformatics is of course a field which is much bigger than just biochemistry. But biochemistry is a very good basis for bioinformatics. So if you are a biochemistry student, I would say learn coding and learn more about biology. So don't just focus on the chemistry part, branch out to things like genetics and transcriptomics because these are the things that will help you become more all-round in the things that you do. Do you have experience in GWAS meta-analysis? Yes, yes. I actually know a lot about QTL analysis and GWAS and I do know how people do GWAS and meta-analysis. The only thing that I don't really like about the whole QTL-GWAS field is that you're only doing statistical analysis, right? So in the end, you get a region of the genome which is associated with your phenotype and then you still have to figure out which gene it is. You still have to go to the lab, validate your findings. So I've published some papers on GWAS. So we have done GWAS in cattle when I was in Berlin, when it comes to milk production and different milk quality trades. When I was still in the hospital, I did human GWAS so when I was still working in Groningen. But the GWAS field is difficult to become good in. But yeah, I do have experience in GWAS and I do a lot of meta-analysis in that sense. And that is actually the only way nowadays that people can do genome-wide associations in humans because you're not going to get a million human phenotypes and a million human genotypes. The only way to get these kinds of sample sizes is by kind of doing meta-analysis across different studies, different cohorts, different fields. If you weren't a bioinformatician, what would you like to be? So I always say if I wouldn't have done bioinformatics, I would have done history and I would have become a history teacher. I love talking about history, how we got where we were and I think that history is one of these fields which teaches you a lot about what's going on currently. So I love history and I always say I would have probably become a history teacher and I'm sticking with that. Do you like the idea of portfolio projects on public data as a part of your Gitte personal website to advertise your skills? Yes, yes. If people send me their CV, it is the first thing that I look at because I am a very code-oriented person. I want my undergraduates and graduate students to code themselves. So if I see someone that sends his CV to me, if they have a nice email, so a personal email written to me as a person, then I check out their CV and I always check out their Gitte, see what they have coded, see how they code and I think that that is one of these things that if you get past the first initial selection, it can make a massive difference in your employability. So if you are able to show that you did nice little interesting projects or if you have a repository that has been used by other people and has like an X number of stars, then this is a really good way to distinguish yourself from all of the other people that do not have that. Because a lot of people, when they graduate their masters, they have done one project, they haven't published any papers. So if you're in that pool, right, then having this publicly visible repository where you show, this is my code, it's all nicely documented, it has a readme file, it has a license, all of these things, it gives you a big plus. So it really helps you become more distinguished than someone who just did a master project, has a bunch of scripts or used a standard tool developed by others. So to be able to show that you can code and have a really nice example repository that is definitely something that will set you apart from the crowd. I am a statistics graduate, but because I live in Turkey, life science and research are not valued. I like to combine statistics and biology. I try to learn by my own means. That is really good. So yeah, Turkey at the moment is of course a little bit difficult because bioinformatics is really done in kind of the West. So I would say that there are some good groups in Turkey working on bioinformatics. I've been working with a Turkish group which does FDR. So that is spectroscopy measurements. So they look at fats and samples using different spectroscopy measures. So there are some good universities in Turkey where you can become a good bioinformatician. So but yeah, learning by your own is something that you will always have to do as a bioinformatician. Thanks for your advice. Can you recommend courses that will qualify me before making masters in bioinformatics? As I mentioned before, I'm a bioinformatic student. So you want courses that give you probably a piece of paper certifying that you followed the course and that you passed it. I find this always really hard. There are some really good courses out there, but those generally are like my courses. So they're just online. You can follow them, but you don't get a certificate of attendance or these kinds of things. So I will think about your question because I do think that there's a lot of people that are looking for some good courses that they can do to become more employable or to become better or more easily hired for PhDs or masters. I might make a video about something like that where I go through and I show different courses. So I would rather not answer those directly, but I will get back to you. Keep an eye on the channel. I might make a short video where I go through some good courses. What do you like to see? Good coding practice, solid science, biology, understanding. What do you not like to see? So yeah, I think good coding practice is one of these things that can distinguish you from other students. So have a good example repository on GitHub where you show this is how I code, right? But this repository, of course, it needs to address a real biological question. So it can't just be a re-implementation of an existing tool, right? So writing blast but worse is of course not going to work. So I like people that have programmed during their master and have done something new or unique. And that in the job interview can actually point towards that saying that, okay, so we did this thing, right? And normally people just use these tools, but I decided to go a slightly different route and then show that you have scripts to back that up that kind of show how you solve this question. What do you not like to see? So what I don't like to see in people that do a job application to me is a standard cover letter. A cover letter where it just says, hey, name, blah, blah, blah, blah, blah, blah, see my CV attached, right? Now, write to the person that you are interested in. Mention my papers. Mention either my YouTube channel or some of the work that I did and tell me why you want to work with me. I've been accepted in a PhD program in microbiology lab in Xion, expecting to start this September. Any idea what kind of research could be done related to bioinformatics. So in microbiology, there's a lot of work being done on antibiotics and on discovering these kind of hidden gene clusters. So a friend of mine developed anti-smash, which is a program for detecting secondary biosynthesis clusters within bacteria. So the secondary clusters generally only activate under very specific circumstances. So they are not active all of the time, but if you stress the bacteria using certain chemicals, they start expressing these genes to deal with that. So I think that there's still a lot of interesting research to be done there to discover how bacterial stress can actually help us discover new enzymes and discover new pathways in that. And of course a big congratulations on getting accepted because the Xion University is a very well-renowned university and it's a very cool city as well. What would be the difference on doing a master degree in biostatistics versus one in bioinformatics? So it depends a little bit on because you have like a lot of these words going around, right? You have bioinformatics, computational biology, biostatistics. For me, it's all the same thing. I don't see a really big difference. The only thing that is biostatistics means that you are specializing more in the statistical sense, right? So you are doing more genome-wide association type of studies. You have a very solid understanding of statistics. So you are not building websites, for example, which hold large amount of data, right? So you're not doing UI work. So I think biostatistics in itself has kind of more of a focus on the statistical part. Well, bioinformatics is a field which is very broad. So you need to know a little bit about everything. I can make a website one day, the next day I spend on doing data analysis, and then the day after I'm writing a project proposal and then the day after I'm writing a paper. But I would say that in that sense, the education that I had was more in the direction of biostatistics, because statistics was the thing that was really new for me, right? I come from a computer science background, so I knew how to design a website. I knew how to code, but to do statistics was new for me. So my own background is more biostatistics in a way than real bioinformatics. Is a bioinformatics degree like a PhD? Will all of the work be dry lab or some wet lab should be done by myself? That is up to you. I would say that the best bioinformaticians are the people who do work in a lab. I try to be in the lab as much as possible. If there are people that have issues, or if they're just a hand short, I will volunteer. I will go and help them in the lab because it teaches you about the analysis, right? You can see how data was generated by being in the lab, and you understand the weaknesses, you understand the strengths of what they are doing. So it makes you just a better bioinformatician by being in the lab as much as possible. But generally bioinformatics masters are dry lab. Bioinformatics PhDs generally have a wet lab component. So they generally have a component where you have to extract DNA or extract RNA, send it in for sequencing, and then do deep downstream analysis. So a PhD is generally a combination of both. But since it is bioinformatics, the main focus will always be on the analysis of the data, not on the gathering of the data part. What's Oscar doing right now? I have no idea. I'm sitting in my office, so I can't see the cat. I would bet that he's sleeping somewhere. So he's probably on the bed. Oh, my girlfriend answers. He's staring angrily at my girlfriend because he has not been fat yet. So that happens, that happens. Thank you so much for your support and inspiration. I want to be a lifelong learner. You're welcome. I strive to be a lifelong learner as well. I learn new stuff every day. I enjoy learning new stuff. And the day that I stop learning new stuff is probably the day that they can put me in a retirement home. Go order him a pineapple pizza, Anna. No pineapples on pizza. Our cat Simba starts climbing on stuff if he isn't fat. Yeah, Oscar is very verbal. So as soon as he doesn't get fat, he starts verbally telling you that something is wrong. So, Deona, you can also study cryptic genes. Yes, cryptic genes is also still a very hot topic in bacteria. So cryptic genes are genes which are generally not expressed or are not expressible anymore. So those are genes which lost over the course of evolution their promoter. So people start putting promoters in front of these gene sequences and then study the protein that comes out. Very interesting topic in microbiology as well. Thank you, Julia. I think the difference between bio-studs and bio-info would depend heavily on the lab and the person gets into, yeah, yeah, yeah. I think that bioinformatics and bio-statistics, they are in the overlap of the same Venn diagram. If you draw a Venn diagram of what a bioinformatician does and what a bio-statistician does, then 90% of the Venn diagrams would overlap because there is a big, big overlap between the two fields. And like I said, I don't really believe in putting these labels on things. I think bioinformatics, computational biology, they're all kind of the same thing. They're just a way for people to kind of show their speciality in a certain area, right? Someone who says I'm a bioinformatician is generally more all-round than someone who says I do bio-statistics because they will be more on the statistical part. While someone who says I am a bio-data analyst is someone who's really good at working with big data and building infrastructure to hold big data or in the analysis of big data. So I think that they're all similar sides of the same coin. I have done G was in lung cancer and found two methylation QTLs. I have to write my master thesis, how can I extend this result? So the result in that sense is the association. So you found these associations, right? So you found these QTLs for methylation. So the next thing that you could start looking at is look at which genes are there. You could look and see if there's data, which is available on high C, to kind of get a look into how the 3D genomic structure is there and if these MEQTLs might have a modifying effect on the structure. Look at things like histone tracks and try to kind of build a story around the two regions that you find because of course the research doesn't end there, right? And in the end you want to couple this back to the phenotype. So you want to couple this back to lung cancer like we now have these genomic regions. So what do we now know about your type of lung cancer that you're studying that we did not know before and how is this going to help us kind of prevent lung cancer or help us cure the lung cancer, right? Because that is always the goal. The goal is not to do the analysis. The goal is to help people or to find something new which then can be translated into either prevention or a cure. Did an intro to BioSTATS course a few years ago and the professor focused so much on the programming side and less on the actual application of statistics and biology. Yeah, because that depends a little bit on the background that people see coming in, right? In my courses they're also very focused on or at least the R course is very focused on programming, right? Because it's a programming course for people who have not programmed before. So it depends on how people get into or what kind of entry level the students have. So a BioSTATS course if people coming in are biologists then it will focus much more on programming and statistics than when you follow a biostatistics course at a computer science department, right? Because there the focus will probably be more on the translation of the biological problem into a statistical problem that you can test. So it depends a little bit. But BioSTATS is of course statistics. So it is logical that they focus kind of on the programming side. But in the end, never forget that research is done for answering the original kind of question, right? To validate your hypothesis saying that, yes, there is a genetic component and by understanding the genetic architecture we can now say this or we can now do this and we could not do this before. Do you think JET GPT-V will affect bioinformatics work in a negative way? I actually think JET GPT will actually be a big boost to bioinformatics and I'd rather not say too much about it because it's still very new and I've been observing it with great, great interest but I am a very, I would say, conservative person in that sense and I don't see me losing my job to a large language model anytime soon but I do see the usefulness because if you are a bioinformatician who is using JET GPT you have a massive advantage compared to a bioinformatician who is not using JET GPT. The problem with JET GPT and the things that people are trying to do currently with it is that it is a large language model. So I always describe them like this. It is like this mirror in Harry Potter. You look into the mirror and it shows you what you desire most. A large language model is the same. You prompt it, right? So you give it words or you give it a sentence or you give it multiple sentences and it will do what you desire most based on what you gave it but it has no real knowledge in there. It does not hold truth. It does not know what a true statement is. It doesn't know what a false statement is. So it is a very, very good tool but you still need to evaluate yourself if the thing that it's telling you or the code that it's giving you makes sense for the thing that you are wanting to do. So I don't think it will affect it negatively. I think it will make everyone more productive so start learning about it but do not worry about it. That's my advice. I just wanted to say thanks again been worried about my future in bioinformatics and now I feel like I have some direction and things to do and then aim for after this AMA. Thanks. I'm happy. I'm glad. I think it's amazing that me just doing these kinds of things and making these videos is inspiring people to do more. That's like as a kind of teacher because I always see myself as someone who does research but teaches kind of half of the time. It's always so nice to hear people get inspired and want to do more. So you're welcome. What do you think German universities focus more on as mission requirements for masters in bioinformatics? So there is a big difference between German universities where I used to be and between UK universities. So UK universities are a very much for profit system. So if you hit the minimum requirements they will generally let you in because you're paying a fee. German universities are much less motivated by money that the students bring in. But let me re-read your question. What do you think German focus more on as mission requirements for masters in bioinformatics? So this really also depends where you want to do your masters in bioinformatics. If you want to do a master in bioinformatics at a biology department they will generally focus very heavily on having done a biology undergraduate and having done a project in your undergraduate. While if you follow a bioinformatics master at a computer science or kind of IT then they will focus much more on your programming skills that you have obtained during your undergraduate. So it really depends which branch you go into. So if you're applying to a bioinformatics master at a university which is primarily IT focused or at a department or not so much a department but at a university which is very focused on biology the entry level requirements will be different because generally the biology will demand that you have spend time in the lab and no biology while the bioinformatics lectures or the bioinformatics master at a university in a department of computer science will generally focus on are you able to program, have you programmed? Do you have a profile already on Github? What are your computer skills? What are your coding skills? Because we are going to more or less bring in the biology part into your program. Is chat GPT why you started Python NLP course? Yeah, yeah, I not just chat GPT but I'm always interested in new kind of machine learning techniques and up until like two years ago I would have told you that all machine learning is nothing more than automated iterative regression analysis which it still is. But the emergent properties that we see for these AI and machine learning kind of big large language networks they are interesting. So that is why I wanted to know more about the transformer architecture and about having multi-node attention and these kinds of things because I think these systems might be useful to the work that I'm doing as well. So I just wanted to know more. The thing that bugs me the most is that I have not yet been able to find a good example without using all of these Python libraries. So a lot of this magic is in these Python libraries and the problem there is that I kind of want to go one level deeper, right? I want someone to do a large language models from scratch kind of lecture like my RNA sac from scratch lecture. So I'm kind of looking for something like that but I haven't found it yet. If I am unable to find something like that and I have enough time I probably am thinking about doing something like that for the channel because AI is completely hip at the moment. One of the things I like the most about Bioinformer is the versatility and potential in getting involved in different research area as your publications show. Can you elaborate on that? Yeah, sure. I think that as a general bioinformatician I always try to help people with their data. A lot of biologists do very, very interesting work. Like if I think about my collaborators like people like ZHUM, they go to Sudan they spend a lot of time collecting all kinds of phenotype data like hair and blood samples on goats there and then they come back they do all of the work in the lab but they don't have the skills to really in depth analyze it. So that is where I come in and I actually get the fun part I actually get to make the discovery I get the data, I look at the data and I am the first one to see problems. I am the first one to see other issues but I'm also the first one to see the results and that is I think what kind of drives me in bioinformatics is that in a way you're like a service because you are helping people analyze their data but on the other hand you do get to have the fun stuff. You don't have to spend all of these hours in the lab but you do get to make the discovery in a way so you do see oh it's this gene oh see that's what we're looking for so I'm really really interested in that's something that really drives me and the nice thing about bioinformatics is is that if you have a skill like being able to program, doing statistics then you can work on all kinds of projects for me it does not really matter which animal I look at it can be a goat, it can be a cow it can be a human, it can be a mouse it can be a plant so my publications literally are all over the place I've written publications on Arbidopsis about Brasica I've written publications about mice about goats, about humans so yeah I think that that's one of the nice things you never know who is going to be your next collaboration partner I'm currently talking to some people here at Northumbria who actually do ancient DNA so they are looking at kind of reconstructing environments from DNA that they get out of caves so I'm very interested in seeing if I can actually contribute to that and for me it's a completely new field so it means learning new stuff and just expanding my current skill set hello professor good to see you after a long lapse what are the core biophobetation skill you recommend for the area of plant breeding so if you want to do plant breeding be good at statistics be good at programming but never forget sight of the real goal so the real goal for plant breeding is to create more yield and of course there are many different ways to create more yield but one of the things that I like about plant breeding is to look into adaptation of plants to different environments so if you actually can make plants grow in more saline conditions then this will open up agricultural areas for this plant which you did not have before so instead of just focusing on yield as a single kind of number this many tomatoes from this size of a field also think about the larger problem of agriculture that we have that many areas of agriculture are currently not available for the types of plant that we are growing so I think that that is always interesting but be excellent at programming be excellent at statistics because there is a lot of statistics in plant breeding to find the regions of the genome associated with it but also be a critical thinker about how does agriculture work because a lot of people have a very weird idea on how plant breeding and how animal breeding kind of works and focus on the reality of how the industry is nowadays and be able to translate your findings into real things that people working in the industry can use where was I, where was I specifically which programming language do you prefer so I always tell people that do not know any programming language to learn R first but I think R is a really, really good programming language so you should definitely learn R after you have learned R and become proficient at it you can either start learning something like Python however there is something to say for getting a real programming language under your belt like C++ C++ can be combined with R and it will help you do analysis faster much faster than Python can or R can so RC or RC++ is a really good combination R and Python is a really good combination as well so can you tell us about some challenges you have encountered on your career like big difficult data or learning new tools some challenges I think the main challenge that I've been faced is the fact that as a bioinformatician it is really hard in an academic career to get first or last authorship and this is due to the fact that you have a skill and this skill is not valued enough by the people that do biology because the people that do biology they spend hours in the lab extracting DNA or going into the field and collecting their samples so they just see you sitting behind a computer analyzing their data and they think that because you do the analysis within a couple of hours that your contribution is not as great as theirs and that is not true because I literally spend like eight years or more or less my whole life like I've been programming since I was four years old so I've been literally spending all of my free time on learning these skills to be able to do within hours which what if they would have been needed to do it would have cost them like probably like a couple of weeks or perhaps months of analysis time but because of that they undervalue your skill and that is so as a bioinformatician one of the main things that I always run into is that people don't view your contribution as being equal to theirs so it is always really hard to get first-authorships in an academic career Code Academy has an NLP course I think I'll figure it out Dear Professor Adams, thank you from the bottom of my heart for dedicating your precious time for helping poor students like us I truly appreciate your heartfelt kindness You're welcome I'm here to serve you Thank you for spending your precious time to look at me ramble on a Sunday afternoon I learned your course about bioinformatics and I have quite a strong background about them Is it necessary for me to study Python or is R enough Professor Adams? I would say you could probably go your whole career without touching Python I sincerely detest Python as a language I come from a C slash C plus plus background I like curly brackets I like putting dot commas at the end of statements and Python just gives me a headache when I look at it and every time my cat runs across the keyboard and inserts a tab all meaning of code changes I hate that but you should be able to use it and debug it because there's a lot of tools which are available in Python which are not available in R so you should be able to learn at least a minimum amount of Python so you can call out to Python libraries when you need them I learned your courses about bioinformatics it's the same so yeah I would say R is enough but do learn how to do the minimum in Python so be able to use other libraries the real goal of plant breeding is to make the plant breeder money and that is a very sarcastic way of looking at it sure in the end everything is about money but money is nothing more than a currency that we use to value each other's time so with money you can buy time of other people's work life so if you work for 40 hours a week then every hour that you work is just a simple monetary unit so it is in our society how we value things and I think that money in itself should never be a goal but money is a proxy for the things that we value I'm not interested in scientific research and publications and at the same time I'm interested in bioinformatics is that acceptable? yes, that is perfectly acceptable there are a lot of really really good bioinformaticians who do not work in academia they work in research labs for either big companies like Philips or Amazon or Google and there are some really really good bioinformaticians working at startup labs which also do not aim to get publications or academic research but I would say that in my opinion or from my perspective I think that that is a shame because as a bioinformatician you should always try to do science in a way be it just basic research I've seen people or I've trained people even in the past that got on a PhD in bioinformatics and then they just work at a company where they just do production jobs and that is such a shame because if you really have skills and you are good at what you do then doing the same thing every day might make you a lot more money than people who stay in science but in the end the fulfillment I think is greater from doing scientific research and really trying to push human knowledge forward instead of just working at a company 9 to 5 doing the same trick over and over again just for a different client but you can still do that in bioinformatics you can still work as a one-trick pony and do nothing but DNA sequence analysis how AI bioinformatics can be used to look at host microbiome interactions so I think that that is a very very broad question so how microbiome interactions are interactions between where the genetics of the host and the genetics of the microbiome or the composition of the microbiome change a phenotype so in that sense I would always say that if you look at a problem like this like what is the goal to understand phenotypic variation so phenotypic variation of an individual is driven by partly genetics environment and environment part of this environment that you are in is your microbiome so statistically speaking microbiome in my mind should be like a gene times environment interaction in a kind of model so you model the genetics you model the environment and then you have the interaction between the genetics and the environment and microbiome in my mind is just environment it is just measurements very precise measurements of environmental conditions the big problem that I currently have with the whole microbiome field or it's not so much with the whole microbiome field a lot of people are doing really good work of course but the microbiome is something that is changing depending on which hour of the day it is your microbiome changes if you measure microbiome just after eating and then like five hours later you will probably see variation in the microbiome so because of this variability I don't think that we have a good handle on this variability yet to really do a lot of good statistics on it but a lot of people are trying to do this and it's a very interesting research field so it is definitely something where bioinformatics and statistics is already used a lot bioinformatics is used to to store the data to do the analysis you use statistics how AI can fit in there it depends on how you view AI but I don't think that a large language model is going to help you here hello professor do you have a community based group where you discuss about bioinformatics related stuff and doubts poo no I have my well I have my YouTube channel here I set up a Facebook group for my YouTube channel but the problem is that I already spend a lot of time on a lot of different things and running a community is very very difficult so I would say that my community where I generally go to for bioinformatics is on Reddit so the bioinformatics subreddit there's a lot of good people there and you can get help with the questions that you're doing over there all right Solomon my lab just completed a project where we made broccoli that can tolerate higher temperatures that way we can plant and grow broccoli for longer during the year see that's what I'm saying like it's not just about making the broccoli bigger making broccoli grow in places where it didn't grow before is just as much as a net gain as well Dear Professor Alans would you mind if I send you my CV I'm a wet lab student but I actually want to apply to your lab you're a very great professor sure just drop a CV in my email box I am currently not hiring anyone since I don't have funding at the moment but if you are a self-funded PhD student or if you're interested in self-funded PhD then we can discuss something like that and always send me a CV because I keep them all or at least the ones that I think are interesting so if in the future something comes up I always look through the list of CVs that I have so I can match people to projects set that cause a company paid my lab to do the broccoli research we did to work they get the crops which cannot be grown by others cause they are male sterile and that's where I have the issue with so when companies get involved companies want their IP companies don't do things for the benefit of humanity but in the end it's the way that currently our society works the entire concept of companies like Monsanto just highlights the heavy focus on money and less to the goal of feeding people I definitely agree and that's something that is an issue in our society but it's not an issue per se with bioinformatics which do you prefer Python or R so I prefer R a lot of people I know prefer Python while learning C or C++ be helpful for the area of plant breeding in comparison to R and Python so in plant breeding we nowadays use a lot of algorithms like restricted maximum likelihood or very complex generalized linear models which if implemented in R are very very slow or even if you would implement them in Python really really slow so having a language like C or C++ which is really really fast at crunching numbers I don't think that anything can beat the speed of C because C is kind of the closest to the metal that you can get of a processor so I think that C or C++ is going to give you a really really nice spread because you can program the algorithms in R if you then hit issues with the algorithms running too slow you can look at your code you can see or you can profile it you can figure out what part of the analysis is running slowest and then you rewrite that part in C or C++ and then in the end you get a massive speed up and C can just do it much much faster than R can so that is the advantage of C or C++ is the ability to do real real number crunching in a way that R is unable to Dear Professor Alans could you tell me which essential virtues or characteristics of a PhD candidate that you think he she should have I think someone should be honest in the topic that they are researching I think they should be honest and I think that they should be hard working kind of that's kind of the three but I don't really have any like the essential virtues is that people don't try to cheat because that is one of the things that I really can stand people who try to take shortcuts saying that I don't want to learn this I'm just going to use this package which someone else wrote no I want people who are genuinely interested in the topic that they are studying and that are genuinely interested and willing to put in the time to make new discoveries and drive the field forward that's true Solomon but from a research perspective is that there is a need for for more ethical sustainable goals like providing better yield in harsh environments or culminating climate change yes and I would say that one of the one of the other things about plant breeding is that there is a real need for sustainable and ethical plant breeding right if you look at countries in Africa there is kind of a bottom up kind of approach where the communities of people that currently do plant breeding and animal breeding they need to come together and from those communities they can build up so I'm not that negative that it's all about the money but we are especially currently with massive the industry is not the industry is pretty gnarly in some places will this life be available on YouTube channel later yes yes when we stop I'm just going to keep it live and for some reason we still have like the half finished drawing of my bees over there but yeah it will be live we will keep the chat here as well so then you can just see the chat and see me responding to the questions great live need more of this kind thank you yeah it's something new that we're trying out or that I tried out and I really think that it goes really well I'm really surprised by the amount of questions that you guys have all right so if you guys like this and I think you do then I think we should make this kind of a monthly thing or a yeah something that we do more often which tool would you recommend to study RNA splicing so that is difficult but if you just want to study RNA splicing then I would say that there are a lot of default pipelines for that so I can take a look I'm not too familiar with looking at RNA splicing but I would say that a lot of the pipelines that are out there they can take your GTF file with the structure of the gene and the different gene variants that there are and if you have DNA sequencing data they can actually allow you to do calling based on different variants but to look into splicing specifically especially like discovering novel splicing I wouldn't know off the top of my head about any tools all right then 2 hours 15 minutes I think we did really well thank you guys for joining I enjoyed myself tremendously so I will plan another one probably in around a month's time so let's try to have like the last Sunday of the month for kind of an open like 2 hour 30 minutes question hour so thank you guys for watching thank you guys for being here and if you want to see me Played War Fortress I might stream a little bit on Twitch tomorrow so perhaps see you then tomorrow and I will see you guys next time so yeah you guys have a great Sunday and thanks so much for being here see you guys next time so thanks and