Tree to image?

[John Burley]

Hi Will, 

I'm wondering how to create an image of my new phylogeny from the .tre file made by phyloGenerator. Is there a different program to use, or am I missing something?

John

[Will Pearse]

Hello,

Thanks for getting in touch - you're almost done!

There are lots of options for how to view a phylogeny, and you do need a separate program. Something like FigTree (http://tree.bio.ed.ac.uk/software/figtree/) will let you just go file->open and select your .tre file. If you're an R aficionado, something like:

install.packages("ape")

require(ape)

tree <- read.tree("wherever.your.tree.is.tre")

plot(tree)

...will do it for you.

Cheers,

Will

[John Burley]

Thanks, That was easy with Figtree.

That was a test tree I made with a portion of species and genes I'll be using. I did it roughly and with no constraint tree, so am not surprised that it doesn't look great at this stage.

For my next attempt, I have some questions.

1) In the alignment checking phase, I have warnings for 2 of 3 genes. Some sequences for some species clearly need trimming. I am not sure whether to go back to the TRIM stage, or proceed and use 'trimAl' after having done the alignments. I didn't trim in the previous stage because it would delete sequences when I tried this. In a separate post you suggested that this may be because the sequences are no good, although: 1- this also happened on one of tutorials (Silwood), 2- I suspect the sequences are OK because many have been used to create published phylogenies. When I tried the 'trimAl' option, it got rid of the warnings, but should I still have TRIMmed prior to alignment?

That's the main concern at the moment. What I am trying to get out of this is a standardised branch-length value and some measure of dissimilarity for each species so I can incorporate evolutionary distances into generalised dissimilarity models of community turnover in space. 

Cheers,

John

[Will Pearse]

Hello,

Thanks for getting in touch.

It doesn't surprise me that some of the sequences you downloaded in the example didn't turn out to be so great; I wrote the tutorial over a year ago, and the stuff on GenBank has changed considerably during that time. The various reload mode options (target gene length, etc.) should help in cases where you can't seem to find a good sequence.

When you say that some of your sequences have been used to create published phylogenies, does that mean you've loaded existing sequence data into the program using the '-dna' option? If so, it also doesn't surprise me if pG is having trouble 'finding' genes within those sequences, since there won't be any sequence annotations in your FASTA input files and so pG won't have very much to guide its search for 'good' sequences. If not, then if you're using an expressed gene region, make sure you've set the 'type' of gene that you're working with. If you've done all those things, please send me an email with your sequences (use 'output' in the DNA Download stage) so I can take a look at them and see what's going on.

trimAl removes any regions that don't align very well; whether you want to trim the gene down to just the Open Reading Frame (using the 'trim' method) before you do that is really up to you. In my own work, I don't like to use trimAl (I reason that if the sequences aren't aligning well it reflects something about the underlying data), but that's a personal opinion and probably doesn't do justice to the thought and design that went into trimAl. I can't personally see that a reviewer would get angry that you didn't use 'trim' before you ran trimAl, because trimAl is actively looking for the nicely aligning regions, which using 'trim' should be getting rid of too, so I can't see it being a massive problem.

Finally, the 'warn' flag is not very smart, and is deliberately conservative. Examining you output phylogeny, and your output alignment, is probably going to give you the best judge as to whether you've got some good quality sequences. If you have a 'warn', but the alignment looks alright to you... you're probably OK. Species in the final output phylogeny with extremely long branch lengths likely have dodgy input sequences - you can always run pG again with the output from a previous run and correct those species. There's a table that goes through some common problems when downloading sequences, how to diagnose them, and their solutions, in the paper describing pG (http://onlinelibrary.wiley.com/doi/10.1111/2041-210X.12055/abstract) that you might find helpful.

I hope that's of some help - let me know if it isn't or if something else isn't clear.

Cheers,

Will