Problem with sequences' name in tree output from make_phylogeny.py

[marine landa]

Hi everyone,

I have just installed the new MacQIIME version 1.9.1 and I am playing around with 16S sequences trying to make trees. My sequences are numbered from 1 to 183.

I noticed that my sequences constantly lose their name on my final tree. I tried changing the name on the input file (the one containing the filtered sequences), I tried each of these options :

>seq1
>seq_1
>OTU_1
>OTU1
>1


And whatever I use, I end up with a tree that has 183 branches named "seq_0" to "seq_182", which is obviously not the name in the filtered sequences file.

Is that a bug? Am I passing wrong parameters? Is there some nomenclature I have to respect no matter what for the script not to change the name in the final file?


Thanks a lot for your help,


Marine

[Colin Brislawn]

Hello Marine,

Can you tell me all the scripts you are using, in order, that you use to turn your 183 reads into a tree? I should be able to identify at which step the names are changed, and how you can get around that.

Colin

[marine landa]

Hi Colin,

Of course! Thank you for your help.

For this one, for instance, I did an aligmment with pynast :

$ align_seqs.py -i All_177_16S_seqs_headrerChanged.txt -m pynast -a uclust -e 10 -p 0.10 -o aligned_pynast_uclust/

The file in aligned_pynast_uclust/ has the right headers.

Then I filtered the alignment :

$ filter_alignment.py -i aligned_pynast_uclust/All_177_16S_seqs_headrerChanged_aligned.fasta -o filtered_pynast_uclust_g0.90/ -g 0.90

And again, the filtered file has the right headers.

Then I built a tree :

make_phylogeny.py -i filtered_pynast_uclust_g0.90/All_177_16S_seqs_headrerChanged_aligned_pfiltered.fasta -t clustalw -o pynast_uclust_clustalw.tre


I opened the pynast_uclust_clustalw.tre tree file with FigTree (version 1.4.2 which I believe is the most recent) and that's when the names are different. I've used this tree vizualisation program before and never had the problem. I also have run this command in QIIME dozens of times before (with previous versions of MacQIIME, though, this is my first attemps with version 1.9.1) and it was always fine. I did use to run fasttree and this time I am trying the other options (here, clustalw).
Like I said, I tried different names (see my initial post) on the filtered alignment file and still have the issue everytime.

Thanks for your thoughts! I really appreciate it.

Marine

[Tony Walters]

Marine, does the tree created have the renamed tips with a different tree building software, like FastTree? For any tree, although it isn't the prettiest to look at, you can open the .tre file in a plain text editor to see how the tips are named, to make sure that there isn't any issue with the tree visualization software.

--

---
You received this message because you are subscribed to the Google Groups "Qiime Forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to qiime-forum...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[marine landa]

Hi Tony!

Thanks for the tip, I didn't know you could open these files in a text editor.

This is what my file looks like :

((((((((((((((((('seq_0':0.00106,'seq_1':0.0033):0.10662,('seq_75':0.02575,'seq_80':0.03578):0.05648):0.01322,((((((((('seq_107':0.00239,'seq_108':0.00228):0.00286,'seq_98':0.00414):0.00042


(first few items)


It looks like it is the file and not the software (I thought about it but since I never had any issues with it before...).
I'll try and build the tree with a different program and let you know.

Thank you!

Marine

[marine landa]

Hi again,

I just ran the script again using -t fasttree instead of clustalw. This time, the tree has the right names. It looks like for some reason, if I change the tree building method from the default setting (?), I lose this information.

Is there a way to fix this? I really don't want to play "guess who is who" on this tree, ahah.


Thanks a lot for your help!

Marine

[Tony Walters]

Hello Marine,

I'm not really sure of a good way to recover the original fasta labels. I am not terribly familiar with the clustalW app controller code, but it looks like it's remapping the sequence labels because they are truncated by clustalw (https://github.com/biocore/burrito-fillings/blob/master/bfillings/clustalw.py#L489).

It would probably take some custom scripting rename that tree tip labels based upon the order of the original sequences in the fasta file that was fed into the alignment, or a very tedious renaming by hand. It may also be worth trying to align the sequences manually with clustalw rather than through QIIME to see if it is possible to avoid the sequence renaming.

-Tony

--

[marine landa]

Hi Tony,


I am a little confused with your previous answer, why do you suggest doing the alignment differently? The alignment works fine (and the names are the right ones), it is really the tree building process that loses this information.

After running the command make_phylogeny.py with the other methods for -t (muscle, clearcut and raxml_v730) just to check, I end up with trees that all have the "wrong" names. It looks like this truncating of the labels happens with all the methods that are not the default one (fasttree).



Marine

[marine landa]

Hi again,


Somebody from my lab tried to run the same commands as I did with an older version of QIIME (1.8) and the names were the right ones. It looks like it is something that came up with the new version (I am working with MacQIIME 1.9.1).


Marine