Script to remove gap-only sites from fasta alignment?

[kmkocot]

Hi all,

I am trying to find a script or a program that I can call in a pipeline that can remove gap-only sites in a multiple sequence alignment. Can you think of anything I can use? I'd also like to
delete the columns at either end of each alignment where there are only 2 or fewer sequences with non-gap characters if you can think of anything that can do that.

Thanks!
Kevin

[maubp]

I can picture how I would solve this using Biopython, provided the whole alignment can be loaded into RAM.

How big are the alignments (number of columns, number of rows)?

[nilshomer]

Hi all,

I am trying to find a script or a program that I can call in a pipeline that can remove gap-only sites in a multiple sequence alignment. Can you think of anything I can use? I'd also like to
delete the columns at either end of each alignment where there are only 2 or fewer sequences with non-gap characters if you can think of anything that can do that.

Thanks!
Kevin

Given your alignments in the SAM format, you can remove all alignments with indels with the C-program "dbamfilter -i". You could also easily parse the CIGAR field in the SAM format for any "I" or "D" operators.

What format are you working with?

[kmkocot]

Hi maubp,

The alignments vary but all have fewer than 30 taxa and fewer than 1000 amino acids.

Thanks,
Kevin

[kmkocot]

Hi maubp,

The alignments vary but all have fewer than 30 taxa and fewer than 1000 amino acids.

Thanks,
Kevin