 part of this lecture. It took a little bit longer than I expected. I was expecting it to do it in three hours, but that's okay. We can just add a little bit more time on top for you guys. So yeah, I just started recording. So I'll come back. Everyone also on YouTube and Moodle, like, favorite, subscribe, and all of these things if you want to of course. Right, so just finish showing off the last slides and then we can do some questions about the lectures from last week. I think it was really hard, which is on purpose a little bit, because if you're not struggling with the assignments, then you're not really learning anything from it. So what do statistics tell us, right? So descriptive statistics are really nice because we can detect things like outliers, do some exploratory data analysis, and decide which model or distribution we need to use to analyze our data. Besides that, of course, hypothesis testing is something which we use to test if a given hypothesis is true based on the data we observe. Final word on project planning, and this is something that I see going wrong a lot. Bacon, thanks for the lecture. Yeah, pick up your kids. See you next week. Thanks for joining us. So for project planning, I always like this quote of Ronald Fisher, and I show it, I think, all of the lectures, not just in the bioinformatics course, but I think I also show the exact same slide in the R course. And that is, and I'm just going to read it, is to consult the statistician after an experiment is finished is often merely to ask him to conduct a postmortem examination. He can perhaps say what the experiment died off. And that is really true. And I see this a lot in science. And with the people that I work with is that they only consult a statistician when they are unable to get their results significant. And that is not the way that it should work, right? So when you are planning to do a project or scientific experiment, you always want to ask a statistician beforehand, you want to know, for example, how many samples should I use? If this is the expected effect size, and this is my variance, and a statistician can help you to kind of beforehand think about what kind of data you're collecting and how you should collect your data. And that is one of the things that will help make a project successful. I once not personally was involved, but I saw a couple of questions where someone did experiments with apple trees, right? And experiments with apple trees are year long experiments, because when you put the seed in the ground, it takes many years before you have a tree. And they had a really good idea because they made all of these recombinant inbred apple trees so that they can do genome wide association, so to find like modifiers on the genome. And in the end, it turned out that they had five trees too little. So they did a 15 year long experiment, collected all of their data. But then when they started doing the analysis and the association of the data, they figured out that if they would have just planted five more trees, their effect would be significant. And now it was only a suggestive fact. And of course, this threshold is kind of arbitrary, right? This one in 20 threshold that we always talk about, but it is a threshold that is rigorously kept to when we do, when we write papers and want to publish results. So definitely consult a statistician before you start doing stuff, because you don't want to run in the situation where you spend half of your PhD collecting data, and then in the final stages when you write your paper and you do your analysis of the data that you collected, you end up with a sample size, which is just a few animals too little. I would say, because Ronald Fisher did this statement in like the 1960s, is that if you are going to do nowadays a big biological experiment, you should also consult a bioinformatician, especially if you're doing things like DNA or RNA sequencing. If you're doing DNA and RNA sequencing, you have to deal with things like data storage, which means that you have to store for every sample that you collected around 60 gigabytes to one terabyte. You have to deal with computational time. The computer takes time to analyze like massive amounts of data, right? So an analysis time for a single sample in a DNA or RNA second experiment can be in the order of 20 to 200 hours. And you have to plan this in your project, because if you only plan the biological parts of your project, but you fail to plan for the bioinformatics part of the analysis, then you make promises which you can't deliver on in the end. Besides that nowadays, if you think about things like animal welfare, a statistician as well as a bioinformatician can calculate for you the things like sample size estimates. So the minimum number of samples that you need to get significant results. And if you have to do like a t-shirts application, then of course that is also really important to not use more animals than you need. So head there, a bioinformatician can help you. And of course get informed about statistics and sample randomization. Had these things can be done by statisticians, but also bioinformaticians generally have a good understanding and a good grasp of statistics, because it's part of our field as well. Good. So those were all the slides for today. So I talked to you guys first about traits, qualitative quantitative traits, Mendelian traits. Why is qualitative traits there again? Mendelian trait and complex traits. So the third point should be complex trait. We talked a lot or I talked a lot about genetics, the origin of genetics. So we talked a little bit about Gregor Mendel, we talked about Thomas Hunt Morgan, his theory of chromosomes, which is still used every day today. We talked about how you set up a cross, right, and figure out which animals are recombinant when you do a two-point cross. I showed you a small example of a two-point cross, a three-point cross when you want to determine the order of genes on the genome. A few words were said about statistical analysis and multiple testing, and a few words about project planning. So that's it, what I wanted to discuss you guys, but with you guys today. We still have like half an hour left in the lecture, so if there are any questions for today, oh, I forgot to update Moodle. I was so busy this morning, let me actually put the assignments on Moodle, so that you can at least get the assignments. Again, I will upload the lectures probably somewhere tomorrow. I'm having a little issue with the recordings being too big, so I have to recode them, which takes around half an hour per lecture. So let me upload the assignments for today on Moodle. X, we are doing the bioinformatics course, then this is assignments two. Let me see if everything is up to date, kind of, and I will just put them online. So I will say turn editing on at a label, and the label will be, no, don't record audio, the label will be lecture two, phenotypes, and save the assignments so that they are there for you guys. All right, so turn editing off, so if you go to Moodle you should see the assignments now. If there's any questions, feel free to ask them now, and if you just want to talk about stuff, we can do that as well. If you want to see more Meerschweinchen, then you can also arrange that. Although I don't have more Meerschweinchen than the ones that I showed you already, but we can do that. But for the people watching on Moodle, I will stop here, so stop my recording. For the people watching on stream, feel free to hang around and just chat with me and ask any questions that you want. Remember that I'm here for you guys, right? So you make or break the course. I can just be here, do my slides, and no interaction just means that it's less fun. All right, so for people watching on Moodle, see you next week, and that's it.