I have a large concatenated protein multiple sequence alignment. One of the species have many undetermined 'X' amino acid. Such column will probably not add much phylogenetic signal so I want to remove all such column (that contain any X) from my alignment file. I tried to do it with biopython but could not be succeeded.
Recreate the alignment with the software that you used before but without the offending protein in the collection of sequences to be aligned.
Apart from Gerard's great suggestion, you might want to take a look at Jalview. It's a truly wonderful program and it will allow you to remove entire proteins from an alignment.
Just make sure your system has, and then assign, enough RAM to run the software efficiently with large alignments. It's java based so will run on any system.
http://www.jalview.org/
AliView (http://www.ormbunkar.se/aliview/) will efficiently handle very large alignments. It is trivial to search for “X” and deleting the relevant columns. You’ll have to keep track of the position numbering though…
Coding a script in BioPython can solve the problem, and will be a robust solution in case you need to carry out the same task for many alignments, but you'll have to "play" a little bit with coding to achieve that.
The program MacVector will allow you to do this. Unfortunately, it's not a free program and is only available on Mac.
Thanks but I use Linux and Windows. I just attached one of the example alignment. Here I have several X in seq1 and seq4. So I want to remove all columns containing X.
You may simply use MEGA 5 or 6 software in which the aligned file can be edited
>but I use Linux
Unless the FASTA names of your proteins contain "X":
sed 's/X/-/g' myfile.fasta > myfile_2.fasta
and then "Remove gap-only columns" in your alignment editor.
I know this @ Rolf but it is not the good solution to treat gap and undetermined character equivalently for phylogenetics
Assuming you have knowledge on how to execute a perl-script in a *nix environment - Try this, wrote it quickly, so no guarantees (Should work though...):
#!/usr/bin/perl -w
use strict;
my $inFile = 'example_seq.fasta';
my @headers = ();
my @sequences = ();
my $index = 0;
open(IN,'/ ){
$headers[$index] = $line;
$index++;
}
else{
$sequences[$index-1] .= $line;
}
}
close(IN);
my %lookup = ();
for(my $i=0;$i
You may try:
#!/bin/bash
split -l8 example_seq.fasta example_seq.fasta_ |
grep -rlI "X" * | xargs -I{} rm -v {}
(It will also remove your executable file if you put this code into one). Give split a "-a 3" option if you have more than 99 files. Let me know if this does not work, a.
Ashmit:
If you wish to try the snippet I sent you before, please make a copy of your source file into an empty directory and then work on that copy because the snippet will delete any file/subfile containing "X". Thanks, a.
@Artur: The X came from gene prediction with Augustus. I also posted a new thread to know how the researcher treats these cases.
https://www.researchgate.net/post/How_do_you_treat_X_or_N_character_in_multiple_sequence_alignment_for_phylogenetics?_tpcectx=qa_overview_asked&_trid=55316b18d685cc0d528b45f9_
I think the best software. for this can be bio-edit or/and edit-seq edit
@Ashmit Singh : Did you try the script I wrote you? It should do exactly what you asked for?
Using vi or vim (e.g. vi alignment.txt) type ":%s/X/-/g". This command should replace Xes by gaps.
Hello colleague, let me tell you that in the first place you must edit your nt sequence before analyzing (editing and annotation phase), after having the reliable sequences you must search the orf to transform them to amino acids, using paid or free programs : MEGA7 or DNAstar (Lasergene) respectively), once converted to amino acids you must align them using the above programs or other programs like SSE.v1.2 or Bayesian both free).
From needing help in editing your sequences and analysis you can consult me directly, I work with MEGA7 and DNAstar.
Note: I only speak fluent Spanish and Portuguese but my English is basic so I will delay in responding in that language.
Hola colega, dejame decirte que en primer lugar debes editar bien tu secuencia de nt antes de analizar (fase de edición y anotación ), luego de quedarte solo con las secuencias confiables debes buscar las orf de estas para transformarlas a aminoaciodos, para esto puedes utilizar programas gratuitos o programas pagos (Ex: MEGA7 o DNAstar (Lasergene) respectivamente), una ves convertido a aminoacidos debes alinearlos usando los programas antes mencionados ( otros pramas que te pueden servin tambien son SSE.v1.2 o el bayesian ambos gratuitos).
De necesitas ayuda en la edición de tus secuencias puedes consultarme.
Nota: solo hablo español y portugués fluido, mi ingles es básico así que demorare en responderte en ese idioma.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics