1
$\begingroup$

I building a robust amino acid phylogeny with adequate robustness support (see previous post here).

This is a brief summary of what I have done:

  1. I performed a BLASTp analysis using a protein sequence from species A as query against the full NCBI protein database,
  2. downloaded all the hits in fasta format and I added the sequences I identified with the query sequence in my target genome, which is closely related to species A.
  3. I then aligned the sequences using the Compute - Alignment - MUSCLE option in Seqotron and I saved my alignment file in Phylip format.
  4. I tried to run this command line in RAxML to build a robust maximum likelihood tree:

raxmlHPC -m PROTGAMMAILG -n output.tre -o outgroup-p 10000 -s ~/Desktop/alignment_file.phy

However, the program showed the following error:

Error, the outgroup name "outgroup-p" you specified can not be found in the alignment, exiting ....

Question how can I build an outgroup for my phylogenetic tree, and how do I include it in my alignment?

I appreciate your help and suggestions!

$\endgroup$
2
  • $\begingroup$ Can you show us the input file please? Personally I would avoid using a '-' in the filename $\endgroup$
    – M__
    Oct 5, 2019 at 9:47
  • $\begingroup$ Sure. The input file is named simply "alignment.phy" to avoid any problems. $\endgroup$
    – Leah
    Oct 7, 2019 at 1:08

2 Answers 2

2
$\begingroup$

I think you have a bug

raxmlHPC -m PROTGAMMAILG -n output.tre -o outgroup -p 10000 -s /Users)/username/Desktop/alignment_file.phy

Try the above, where username is your username. I'll check my codes later. The most important bit is the space between the -p and word outgroup. You can omit the -p 10000 it is only needed for very geeky maximum likelihood and sets the random number stream.

$\endgroup$
6
  • 1
    $\begingroup$ Thank you @Michael G. I tried using this line: raxmlHPC -m PROTGAMMAILG -n output.tre -o outgroup -p 10000 -s /Users/username/Desktop/alignment_file.phy but I get the following error: The file /Users/username/Desktop/alignment_file.phy you want to open for reading does not exist, exiting ... I also tried using raxmlHPC -m PROTGAMMAILG -n output.tre -o outgroup -p 10000 -s /username/Desktop/alignment_file.phy but I got the same error message. As mentioned above, the file name is simply alignment.phy. $\endgroup$
    – Leah
    Oct 7, 2019 at 1:10
  • $\begingroup$ I've set up a chat group here, chat.stackexchange.com/rooms/99546/phylogeny. Please drop by and let me know when you are around. Tomorrow afternoon your time zone would be good. Please make sure you have access to your Linux/Darwin terminal and your data sets $\endgroup$
    – M__
    Oct 7, 2019 at 9:35
  • 1
    $\begingroup$ Thanks a lot @Michael G. I will be there on Tuesday October 8 from 17:00 (CDT) onward. Does that work for you? I can connect earlier if it suits you better. $\endgroup$
    – Leah
    Oct 7, 2019 at 18:18
  • $\begingroup$ Yes earlier, ie before 17:00 please (6 hour time difference). Must leave by 17:00 $\endgroup$
    – M__
    Oct 7, 2019 at 18:40
  • $\begingroup$ I see. How about 9:30 am CDT (UTC -5), which would be 15:30 BST (UTC +1)? Is this better? $\endgroup$
    – Leah
    Oct 7, 2019 at 20:39
2
$\begingroup$

To clear up what may be a possible misunderstanding about RAxML: RAxML (and most/all other maximum likelihood phylogenetic inference programs using reversible models of sequence evolution) explicitly infer unrooted trees. This is because, for a reversible model of sequence evolution, changing the position of the root does not affect the likelihood of the tree. Therefore, as you will see in the RAxML manual, specifying an outgroup is explicitly a display option - it doesn't affect the inference in any way. This means that it's usually easier to just infer the tree in RAxML without specifying any outgroup, and then use any tree viewing software (e.g. FigTree, phylotree.js, or equivalent) to reroot the tree as you need to.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.