Phylogeny based on multiple sequences

Question

[Mehmet]

Dear All:

I have 113 ortholog groups of 12 species. I want to make a phylogenomics tree based on these orthologs. I was wondering how I could do this. After alignment of each ortholog group separately, should I combine them and then do the phylogenomics tree or do phylogenomics tree of each ortholog groups separately and then combine the each ortholog group`s tree?

Thank you.

Answer 1

[Naren]

You can either align them first and merge alignments to one. or concatenate all representative sequences and align them as one sequence per Ort. group. MUSCLE is my choice for alignment as well as tree generation.

[5heikki]

Really fancy way would be to estimate best substitution model for each alignment post gap removal but prior to concatenation and then give that info to RAxML along with the super alignment.

[Mehmet]

Hello:

I aligned each ortholog group one by one and merged them by using "cat" command. but I received an error during trimming. The error says sequences are not at the same length. I want to use alignments in nexus format for MrBayes.

[Naren]

cat will vertically merge them all but you need to merge them horizontally somehow.

[5heikki]

I don't know in what format your alignments are but it's almost certain that you can't merge them with cat..

[yixinrufeng]

could you please tell us how to merge alignment files?

[Mehmet]

I used a script (FasconCAT), which is available at https://www.zfmk.de/en/research/research-centres-and-groups/fasconcat

[eeb109]

Instead of cat, paste command is possible to concatenate the sequences side by side. But you have to make sure that the FASTA sequences are in one-liner format before concatenating them using paste. After pasting, remove the delimiter and you will get your concatenated sequences. Remember to pick the delimiter carefully. Pick delimiter that does not exist in your header or any character that might be in your sequences as. Else, it will be a disaster to remove the delimiter later. Eg.

paste -d' ' seq1 seq2 | sed 's/ //g' >> concatenate.fa

The command assigned space as the delimiter and the sed command removed the spaces after pasting. *There are more better way perhaps, for concatenating sequence in command line. This is for sure not the best one, but it somehow work.