The Joy of Phylogeny: How To Make Your Own Phylogram

[MonotonePeanut]

Thanks! Playing with phylograms is going to be so fun ^^

[jebus6kryst]

Thank you for this!

[jebus6kryst]

HAHAHA!

It is always a hoot to read something from you about a subject you know nothing about. Maybe next post you will inform us about "radio-active" dating methods.

[chaosunstabledynamic]

Just to let you know, if you look at this phylogeny, you'll see that rats are the outgroup to the whole tree- you can't say they pair up with the mice like you did, because the tree is actually saying that the mouse GAPDH gene(s) share a more recent common ancestor with ALL of the other GAPDH genes of the other species (e.g. humans, etc.) than ANY do with rats. This isn't saying that mice aren't close relatives of rats, just that the phylogeny of this gene from these data don't support that.

[C0nc0rdance]

@chaosunstabledynamic

unrooted phylogenies don't really have outgroups.  It's arbitrary which is the reference sequence. This is the least rigorous kind of phylogram, but I wanted to show people how easy it is to generate them without a background in bioinformatics. Just for fun.

For the record, beta-actin is a 100% match between mus and rattus, for 375 amino acids. It doesn't sound like you challenge that relationship, just how it was graphed. Correct me if I' m wrong.

[chaosunstabledynamic]

Correct me if I'm wrong (which may well be the case), but can't you still have outgroups in unrooted trees (they're just not specified a priori)?Just seems to me that in this particular cladogram, the rat sequence data seems to be basal to everything else.Definitely in agreement with you about the relationship between rats and mice- but just not as supported by this particular tree.

[chaosunstabledynamic]

Just not sure you can say what was said around 9:06/9:07, although I think such a comment would be supported by most phylogenies. Sounds like we're mostly in agreement here.

[grayziehobbes]

@chaosunstabledynamic We are a rodent. That does not suck.

[Aneurythmic]

Very interesting, good tutorial.

[FlyingEvolutionist59]

I'm probably the only person in my biology class who appreciates this, but I find genetics and phylogeny (and of course evolution) fascinating. It's great to see introductory videos on formidable forms of professional bioinformatics research.

Thanks very much for the video, it was greatly enjoyed!

[BlahKing101]

Lol@Slapstik in the gene sequencing at the beginning of the video.

Great vid though. ^^

[nemirn]

Funny.

I saw your comment before I watched the video and when it got to that point the SLAPSTIK jumped right out at me.

[latuman]

Fascinating. Though using PubMed etc sites is a daunting task to anyone not studying biology seriously.

I can monkey this guide and find certain gene similarities, but in the end, even though I get great answers, I get even greater questions, and more of them.

Especially, the gene sequences are presented as alphabetical letters. What do these letters represent, how are they assigned etc? Any links that might explain this?

[magick205]

@latuman: NCBI also has Molecular Biology of the Cell (4th ed) online as well. That should give up the conversion table you're looking for. The letters equate to specific amino acids.

[BluMoanin]

Thanks for bringing to light something I didn't realize was so easily available to the public!

[Linux7Master]

This is awesome! Thanks man!

[SauronsEye]

What a useless video....who can read what you've posted?

Oh and from what you were saying...what a wanker.

[pikechris1]

very much appreciated, C0nc0rdance! thanks!

[Minttzz]

Excellent! 5* & Fav

[majorvoltage]

It seems like every gene that was plotted showed zebrafish and chickens as being closer to invertebrates than to mammals. Why?

The leptin receptor phylogeny at the beginning of the video shows fish and chickens being closer to each other than to the mammals. OTX2 has humans and zebrafish closer together than to chickens.

c-Src again has fish and chickens closer to invertebrates than mammals. Same with GULO and tinman. What's going on?

[phoenixshade3]

What you are seeing is UNROOTED phylogenetic trees. I haven't used the program c0nc0rdance is using (I use a different program called phylip), so I haven't checked out any of the options, but most phylogram programs have the ability to define one of the sequences as an outgroup, which thus roots the remainder of the tree.

(more)

[phoenixshade3]

(2/3)

What's important is to focus on the HORIZONTAL distances between organisms. For example, the Rat and Mouse appear to be quite distant from us in the unrooted tree, but notice that the horizontal distances are very short, indicating a fairly close relationship -- closer, in fact, than the fruit fly is to the mosquito.

(more)

[phoenixshade3]

(3/3)

With a little practice you can visualize how the tree would appear if it was rooted using the malaria organism as an outgroup. The first division would set up two clades: one including the vertebrates, insects, and roundworms; the second including the plants, fungi, and yeasts. On the animal side, the next division would include all the vertebrates on one branch and the invertebrates on the other.

[phoenixshade3]

(4/3 ... I lied)

What you are noticing in the differences that you list is basically "statistical noise"... looking at the peptide sequence of a single program is a ridiculously small sample size, and as you probably know, the smaller the sample, the greater the statistical noise. If DNA sequences are used rather than just peptide sequences, a more accurate picture emerges... and if you use SEVERAL aligned DNA sequences (for example, the entire mtDNA), the results are better still.

[phoenixshade3]

"Single program" should read "single PROTEIN."

"SEVERAL aligned DNA" should read "LARGE aligned DNA."

I should really proofread before posting.

[majorvoltage]

Ooooooh ok so if I understand you correctly the relative horizontal displacement represents the amount of difference between the sequences. I remember Dawkins talking about rooted vs. unrooted trees in The Ancestor's Tale but I was a bit fuzzy in visualizing it. I'll play around with it and see if I can find an option for setting an outgroup.

Thank you for clearing that up phoenixshade3.

[aj566]

i had a question if anyone could help. why canis lupus and bos taurus have more sequence identity with human than mus musculus and rattus norvegicus do? mus and rattus are closer to us isnt it? as they and we are both euarchontoglires whereas we and bos or canis are eutheria. eutheria is a clade prior to euarchontoglires isnt it? can anyone explain?

[majorvoltage]

WTF?! Again, in your PAX6 example at 0:15, the zebrafish is closer to invertebrates than to mammals?!

[majorvoltage]

Why is the zebrafish closer to the invertebrates than to other vertebrates, and the chicken closer to the zebrafish/invertebrate group than to the mammals?

[DonExodus2]

Why the hate on the lolcats :(

[C0nc0rdance]

Because, like you, I use them so often to add a little levity to my vids. This one REALLY needed it, but I had a hard time finding the right one to express the concept of "phylogeny".

[VoteNixon2008]

don't fungi have different codons? is the estimation of homology take this into account in it's estimate?

[Laserturkey]

S cerevisiae is also used in making rum isn't it? <3 rum :D

[Kattarina98]

Is it true that beer yeast does not occur in the wild - that it is a culture yeast only? - I am really stupid, as far as biology is concerned, otherwise I might find that out using your method, there should be a way to show how old this yeast is.

Thank you for sharing, ***** as always.

[C0nc0rdance]

We have only documented 1% of the yeast species. Saccharomyces cerevisiae, or ale yeast, is a top fermenter, Saccharomyces pastorianus or lager yeast, is a bottom fermenter. S cerevisiae is also used in wine fermenting, and this is the most common yeast in research laboratories.

S cerevisiae is known to grow in the wild... it's a yeast, so it probably grows everywhere, it's more a question of moisture and dispersal.

Yeast shares 23% of its genome with humans. Amazing, no?

[D0ubleHelixFilms]

Could you give me the name of a gene that has a massive list of species similar to that of the one you used last video? Thanks.

[C0nc0rdance]

Check out the video description on the right.

18S rDNA or 16S would be good. APRT, HPRT, catalase, ubiquitin, actin, myosin, histone, chaperonins, MAP kinase, transducin, Stats, cyclinD, JNK, TLRs, NFKappaB, zinc finger proteins, MLH1, GAPDH.

If you produce something interesting, do a screen cap and post it.

[intelligentfalling]

I did one based on LRP5 which directs bone density in humans (at least thats what I understand)

I dont know if its interesting although I did notice that chimps and humans were farther away than I anticipated. lemme know if you got the vid response.

[intelligentfalling]

The human chimp relation was reflected pretty much exactly in sarcoma

[phoenixshade3]

Basically, any of the mitochondrial genes will have been sequenced for MANY species. There are two rRNA genes (5s and 16s), two ATP synthase genes (ATP8 and ATP6), 3 COX genes, 7 NADH genes, a large handful of tRNA genes, and probably one of the most widely sequenced genes of all, CYTB. That should give you a start.

What's really interesting though is the non-conserved sequences, such as the D-Loop, that can be used to estimate divergences within a species, such as the human haplogroups.

[Salmo77]

Tubulines, both Alpha and Beta.

[nhojmabon]

I've compared different strains of influenza A virus to investigate swine flu and I made a really interesting discovery. Viruses have sex, no joke. I noticed the viruses weren't appearing on the right spot on the phylogenic trees I made and so I did someresearch. It turns out that when a cell gets infected with two viruses, the genetic material can get mixed up, so the viruses can swap genes. This process is called 'reassortment'.

[C0nc0rdance]

And that's a key point in why they call it swine flu. Chicken-tropic InflA viruses can weakly infect pigs, and pig-tropic viruses can weakly infect humans. Human tropic virus can infect pigs.

The pigs act as mixing bowls for genes between chicken, pig, and human tropic viruses. This leads to previously unseen antigen epitopes, and super-FLU.

[Dark1777]

Fantastic video. Thank you very much for making it.

[ngarbo00]

awesome videos! I can barely keep up man! possibly because I'm still a little wasty.

[nhojmabon]

THANKYOU THANKYOU THANKYOU!

[ndjarnag]

I am rather surprised that the results seem so predictable and strait forward.

Here you just type in a gene and out pops the expected phylogram.



(From my experience in a chemistry lab, results are often ambiguous and inconclusive.)

Perhaps compare a few phylogram in future videos.

Awesome stuff!

[C0nc0rdance]

I'm just not talking about the subtleties. There's too much material. But, go back and look at the first six slides. Pay attention to the distance back to a common fork. That's how distantly two organisms are related. It's pretty startling, and not always what I expected.

The effect of this data on me is to create an obsession with yeast. They are more diverse than insects, so far as I can tell.

Some yeast make beer, some make pizza. How can you not love them all?

[ndjarnag]

I am just blown away by the fact that you can input a gene and out pops a recognizable phylogram. It's quite familiar even for a novice in biology (subtleties aside).

[cheesothegreat]

Extremely boring video... but I'm sure I can bore you just as much by talking about the efficacy of nuerotrophic factor on degenerated rod outer segments and retinal pigment epithelium =)

Keep up the good work and the good science!

[godhead22]

Do you know why the rats and mice show up a long distance from the other vertebrates?

[superfisto]

Awesome.

5/5!

and if I could I would rate 5/5 for science.

But I have bad news. Anything with more than 3 steps is too complex for the typical creationist.

[leicaholt]

1* for hurting my brain, but 5* for sharing the information. definitely awesome. see you learn something new everyday

[magick205]

@darthman07: I'll bet the NCBI and Ensemble haven't seen this much traffic since they first went on-line!

[GodKillerAtheist]

way over my head for now LOL

[magick205]

@GodKillerAtheist: There are some very good general bioinformatics books out now. The "For Dummies" series even has one! LOL. Also, there are tutorials on the Ensemble website that will show you how to use the tools. Where you go from there......

[GodKillerAtheist]

Bah! I would just look up Ganja and be done with it LOL

[magick205]

Ah, But which strain?? LOL

[GodKillerAtheist]

Good point. Well I better get to the noob area then =P

[A0Refrigerator]

Wow that looks cool.

[welwitschia]

Why are mice and rats so far away? shouldn't they be clumped in the dog/cow/ape cluster?

[RealityMonitor]

Getting Neurospora wrong... Whatever would Beadle and Tatum think?

Also - Although long IDed as "Red bread mold" - It's normally orange and it's natural environment seems to be very wide ranging. They used to think it was semi-tropical, but recently has been found growing wild in temperate forests world wide.

[C0nc0rdance]

*hangs head in shame*

Forgive me Beadle and Tatum.

(for those of you who aren't bio-nerds, Beadle and Tatum developed the one gene-one enzyme hypothesis on Neurospora in the 40s)

I was thinking of Dictostylium.

[TooMuchButtHair]

You can also use BLAST for a great deal of what you did there, but your way is easier (never tried it that way :p).

[TheStephenation]

I'm familiar with cladograms from biology classes and some books I've read, but phylograms I'm a bit unclear on. I can imagine cases where the additional information would be useful, but is it as reliable as the stuff already shown in a cladogram? Are phylograms preferred whenever possible or only used for specialized purposes?

[boonw]

thanks, I knew phlogenetics was based on dna similarities, but i didnt know how the organized it until I saw this video.

[Qieth]

Id wish i knew enough biology to understand this, as it would surely help me understand the world more. But alas, i have nothing to contribute to the video, other than saying job well done! Im sure it will be helpfull to others.

And because we lack them, i felt that i should add a comment that fundies would write:

"Lul, you dont know all this for sure, its just theory, so its stupid to follow it".

Yeah... ... ;)

[simsi1616]

I <3 Genbank!

[tomaselvis]

What's with the phaser effect on the voice???

[ProfetaParanoia]

yeah... I noticed that weird flanger too.

[C0nc0rdance]

Stupid automatic noise reduction on Camtasia Studio. Knocked the volume down, too, which I didn't notice until after I uploaded. It's annoying, but you can understand what I'm saying.

[ProfetaParanoia]

Yeah, just feels fine at moments, but it's a just detail. ;)

I'm really not into this stuff, but was interesting to watch!

[Roriniho]

I just experienced a nerdgasm. Was Wonderful.

[Noobwarriking]

great vid

i myself am an economist(studying)

and i have not much understanding of this but i would love it if you could take us through the typical processes of biologist studies

that would be really exciting

[stefanlittle]

To follow up. You have a pair of two fungi near the plants that are not linked to the other fungi. This is probably due to overall similiarity (ie convergence) of the protein sequence, but ideally a cladistically formed tree should put all the fungi into a clade (because they have common decent).

[C0nc0rdance]

I think I understand. Then, my understanding is that ClustalW2 is a similarity matching heuristic tree-buildier, with iterative tests of parsimony.

It's not meant to be reflective of an organism's true phylogeny, just the best-match similarity tree for a specific protein.

But, I'm just going to call it a phylogenetic tree, so my brain doesn't explode.

There's another tool you would probably prefer at NCBI: BLink. There's a parsimony fast evolution tree builder.

[stefanlittle]

well, sure. But there isn't really parsimony.... but yah it more or less works out either way.

[ScientiaVeritasEtLux]

Homologene is awesome, I love it. I compare all kinds of genes with it like cytochrome c, dna polymerase, helicase, actin, etc. And I agree that it's fun to make phylograms, and it's cool because no matter what protein you choose they always reflect the same basic phylogram (with occasional differences just due to genetic drift I suppose). Great video.

[kintarocrab]

great video, thankyou

[SchrodingersFinch]

Great video!

Can you use the program to make one phylogram based on several different genes?

[C0nc0rdance]

Yes. Give it a try and see what you come up with. The only challenge is finding the sequences you need in the database.

I'd recommend you search for superfamilies in the protein database of NCBI. "HLH AND Homo sapien", for example, or "HOX AND Homo sapien"

Then you just have to sort through the jumble of proteins and fragments, outdated records and odd verbage.

[Crowfighting]

That's deeply cool. I'm forever agog at how technology has brought knowledge and research tools like this within reach of even the keen amateur. One needs only the desire to learn (and helpful professionals to explain it all!)

[stefanlittle]

Oh yah... sometimes there are lateral gene transfers between hosts and parasites... another hypothesis to be tested for why the weird mold thing went with the plants... one could test the lateral gene transfer hypothesis with different shared gene.

[stefanlittle]

Cute. Good job on the video! I couldn't tell (and I don't use that website). What algorithm is being used for the tree construction? Is it UPGMA, is it parsimony? If it is a phyletic neighbour joining type of tree, based on similarity only, then it isn't a true phylogenetic-tree.

[stefanlittle]

I looked it up. Yah it isn't a proper phylogenetic analysis (although many of the times the similarity-based trees do look like cladistics-based tree construction). If you did a parsimony analysis, I suspect that the slime mold won't pop into the plant group. Also, yeasts are not considered to be all closely related; fungi are a really large diverse group that have convergent "body types" yeast is just a general term for blobby single cell/life style, vs filamentous, for example.

[C0nc0rdance]

It actually turned out to be a naming error. What I called "slime_mold" is actually bread mold, which is a classic fungus, so it fit quite nicely.

By default, ClustalW2 uses a NJ (non-joining) algorithm. UPGMA (Unweighted Pair Group Method with Arithmetic mean) is an option.

The NJ does have an iterative parsimony component, as I understand it, and I'd love if you could explain (in simplest terms) why you think Clustal/NJ isn't truly phyletic? The assumption of branching distance?

[stefanlittle]

I'll happily do so once I actually look at the program and see what is says it does... for soem real fun we/I could run the same data in another software program and compare... more soon.

[stefanlittle]

Ok so I did a fast and dirty look up. The programs still essentially rely on similarity and matching. There are improvments in the sense that there are heuristic algorithms that check for global similarity/matching. However, this kind of tree isn't always going to be the same as a cladistics-based tree search (again these can often do look similar). The kinds of trees that are used to infer actual historical realtionships (evolutionary tree) are based on parsimony or other models of change.

[stefanlittle]

parsimony is the simplest to apply the least number of character changes (amino acids in this case) across the tree. Liklihood uses a score based on a gene or protein based model of change, and again there is a heuristic search across "tree space" and the tree(s) with the best scores are kept. It is different than a similarity/matching algorithm. Clustal is more for finding protein groups, not neccessarily actual phylogentic history.

[EvoBiologist]

Great contribution.

Thanks for the plug.

[TheGreatSimonski]

Science!

[Terrencje]

This is one of my favourite subjects in evolutionairy biology (tree of life), I am reading a few papers about the Neomuran hypothesis and the root of the tree of life sitting between Chlorobacteria and the rest.

What I like about it is that the paper uses evolution of flagella, photosynthesis, proteasomes and ecology to name a few to polarise the root of life where Archaea and Eukaryotes are branched deep within the eubacteria, I will be making a video series about it soon.

[Darkasthenight06]

I am going to waste so much time with this.

[pmfcal]

Excellent. Thanks C0nc0rdance.

[TheFallibleFiend]

YT needs more content videos like this.

[pippodeltrappeto]

Science is an orgasm...

[IWontGetOvertheDam]

And I just came.

SCIENCE

*guitar solo*

[snoggerforce76]

wow, really cool stuff!

even without the lolcats. :)

[O2BSoLucky]

I don't understand, I didn't even get the gist. But for people that do, a vid like yours must be great. 5* for the spreading of knowledge!

[Bueller007]

If I'm not mistaken, you said that Neurospora is a slime mold. It's just a regular old bread mold, no?

[C0nc0rdance]

You're absolutely right. It's red bread mold. I was thinking of Dictostylium. I'm not a mycologist... but that's no excuse. My apologies.

[Bueller007]

No need to apologize. I'm not a mold guy either. It's one of the model organisms I recall from intro genetics.

[sinxcosxtanx]

Wow, that is amazing. But, the rodents aren't supposed to be closer than carnivores??

[sinxcosxtanx]

from humans... (oups)

[EdwardHowton]

Fascinating stuff! Definitely a video to keep in mind, and something I'd suggest to any science teacher as a quick teaser for science and biology.

[gj75845]

first, and dosent work :_S