Find an gene as an example of how the human is related to other primates with blast

Question

[bleelii123]

Hi there! I am a student from Germany and working on a research paper about evolution and the human's relation to other primates (bonobo, chimp, gorilla, orang-utan). I have already finished the parts of my paper where i talk about different methods to analyse relations (through morphology, average-distance method etc.) and now I am wondering whether there is a possible way to find a gene, that is in every of the organisms mentioned above and shows after comparing the sequences with the BLAST programme a tree, that fits how the phylogenetic tree looks like ( http://www.nature.com/nrg/journal/v15/n5/fig_tab/nrg3707_F1.html like in this picture, only the upper part). I have already tried alligning sequences of insulin, cytochrom b, s16-rrna and some others but they did not fit. Now I know that a phylogenetic tree based on a single gene is very likely to show a different result than we would expect, but still, does anyone know how i could tackle this problem? The idea behind this use of blast in my paper is to find a way to show other pupils how blast can create phylogenetic trees and it would be perfect to find a gene that suits the right result.

I hope you could understand what i mean :)

Thanks a lot!!

[pld]

Why do you want your phylogeny to match the figure in that paper? Is it even a good idea?

[bleelii123]

I know it sounds odd, but the idea behind this point in my paper is to find an gene e.g. teachers could use in class to show students with blast how everything is 'connected'. The focus is on the methodical-didaktic experience for the students, so they see how working with this programme can look like. Of course a phylogenetic tree based on only one gene has practically no informative value, so I know what you mean.

[pld]

I get that, but I'm not sure why it must match that specific tree. Just show the one from the gene you pick for the exercise. Or, explain that the Nature figure is based on a much more complex analysis taking into account a number of features that a gene-based analysis lacks.

It may not be bad to cover a few examples and to highlight the complexity of estimating phylogeny. It could also highlight how different species have experienced different selection pressures or have responded to similar selection pressures in different ways. Different input information and methods will give different results. It would be bad to give the idea that any gene used will provide that specific result, and anything not matching that specific figure is incorrect.

Using the Ensembl gene trees might be a better approach since the trees are already created, you could just point teachers in that direction.

If you're worried about the complexity of someone not familiar with bioinformatics tools, you could create an input set of sequences to provide with the exercise. EBI has a Clustal Omega web server, and getting a tree from this is matter of a few clicks of the mouse.

[bleelii123]

Actually you convinced me, I'll talk to my teacher about that! Thanks so much :)

Answer 1

[Jean-Karim Heriche]

Blast can't generate phylogenetic trees. It's an algorithm for doing pair-wise local alignments. To build a tree you need multiple sequences all aligned to each other. To get a more accurate alignment of multiple sequences, you should use a multiple sequence alignment program such as Clustal Omega or MUSCLE. From this multiple alignment, you can run a phylogenetic program to obtain a tree.

[bleelii123]

Thank you for your answer! Maybe phylogenetic tree was not the right term, but when i have used e.g. the protein blast and i select a few different results in the table at the bottom, i can click on "distance tree of results" which looks like a phylogenetic tree, is there a great difference?

[Denise CS]

If you want to run a similarity search and find the hits on the human, chimp, gorilla and orangutan genomes, you can try the Ensembl BLAST/BLAT. There is a new tutorial video on it for more info plus this help page. Once you find your hits you can explore all the annotation available in Ensembl including the gene trees and genomic alignments.

[Jean-Karim Heriche]

The blast result tree gives you an idea of the similarity of the hit sequences to your query but doesn't tell you anything about how the sequences relate to each other. You could decide to interpret that tree as a phylogenetic tree by inferring speciation or duplication events for the nodes but this would give you a bad phylogenetic tree because it would be based on low quality alignments (for starters, it depends on your choice of query sequence)

Answer 2

[Ibrahim Tanyalcin]

Dear,

If you pick genes that are essential to cell homeostasis and started evolving long before divergence of primates than you will not get what you want to see. For instance do not try to select genes in glycolysis pathway as every organism you encounter will need a chemically stable version of the enzymes in the pathway.

Try genes that are involved in brain development, especially genes that determine the number of cell layers in the cerebral cortex. When these genes are present with mutations, you will have patients with either under-migrated neurons or less than normal neuronal cell bodies or complete smooth brain with lack of gyration. I didn't try myself but I can give a couple of names for you to try:

LIS1, DCX, NDE1, FLNA, TUBB2B, TUBA1A, CASK, MECP2, ASPM, MCPH1

Lastly try FOXP2 also, I just wondered what that would bring up,

Hope this helps,

Regards,

[Denise CS]

Taking one of @Ibrahim's example, the human FOXP2, we've got a gene tree already for it across dozens of species from Ensembl (the primate branch will be coloured in light purple), from where we got the orthologues of the human FOXP2 in 11 of other primates including all but one of your species of interest (no bonobo). We also have the genomic alignments already pre-calculated for you within 8 primates across the genomic region of that gene with the added value of a tree for those 8 primates only. Both trees and genomic alignments can be downloaded. The later is available in different formats such as Phylip and Mega. Both could be used as input for programs mentioned by @jean-karim I'd have thought.

[bleelii123]

Thank you very much! I am afraid, I could not understand everything as I'm new to this programme. I had a look at the links you've posted and the one with the 8 primates is perfect! My question is now, how did you do that? I am a bit overwhelmed by all the information, so I am again sorry for asking you that if it's obvious.

[Denise CS]

@bleelii123, the genomic alignments across the 8 primates are done by the Ensembl Comparative Genomics team. We've used the EPO (Enredo, Pecan, Ortheus) pipeline for those species. Check out these papers A and B by Paten et al for more details.

[bleelii123]

Thank you very much for your answer! My greatest problem was to find genes that I could try, so this really helps me. And now I see why I should use genes connected with brain development as there is a difference between different primates. I hope that I'll be lucky to find one that fits!