Relaxed and sequential Phylip format conversion

Question

I routinely use AlignIO in BioPython for manipulating alignments and in context here moving from fasta format to phylip format. Alot of phylogeny packages accept a relaxed (sequence ID) & sequential (sequence) phylip format. This format is user-friendly, helps double-check for alignment errors and can be the essential input format.

A typical bit of code is,

'''
Converts a fasta file to phylip format
'''
from Bio import AlignIO

Class Convert (inheritClass):
   
    def fa2phy (self):
        fileOut, fileIn = self._inOut() # builds the path name and deletes any pre-exising outfile
        align = AlignIO.read(fileIn, "fasta") # reads the alignment in fasta
        AlignIO.write(align, fileOut, "phylip-relaxed") # writes in phylip

The output of "phylip-relaxed" is interleaved format

4 120
virusA_detailed_description ACGTGACG .... ATCCCCAA
virusB_detailed_description ACGTGACG .... ATCGGGAA
virusC_detailed_description ACGTGACG .... ATCAAAAA
virusD_detailed_description ACGTGACG .... ATTTTTAA

ACGTGACG .... ATCCCCAA
ACGTGACG .... ATCGGGAA
ACGTGACG .... ATCAAAAA
ACGTGACG .... ATTTTTAA

Output sought is relaxed-sequential

4 120
virusA_detailed_description ACGTGACG .... ATCCCCAAACGTGACG .... ATCCCCAA
virusB_detailed_description ACGTGACG .... ATCGGGAAACGTGACG .... ATCGGGAA
virusC_detailed_description ACGTGACG .... ATCAAAAAACGTGACG .... ATCAAAAA
virusD_detailed_description ACGTGACG .... ATTTTTAAACGTGACG .... ATTTTTAA

There is an option to allow a sequential format (above). HOWEVER, a "phylip-sequential" option output is as follows (it truncates the sequence ID to 10 characters,

4 120
virusA_det ACGTGACG .... ATCCCCAAACGTGACG .... ATCCCCAA
virusB_det ACGTGACG .... ATCGGGAAACGTGACG .... ATCGGGAA
virusC_det ACGTGACG .... ATCAAAAAACGTGACG .... ATCAAAAA
virusD_det ACGTGACG .... ATTTTTAAACGTGACG .... ATTTTTAA

No matter how I look through AlignIO options the above is the nearest format. Could anyone point to a IO style solution? I could write it, but that defeats the object (i.e. purpose not an instance of a method).

---

Note the "...." in the outputs above are not a typical "sequence identity" symbols, it just means that total length of a line in an interleaved format is 60 nucleotides (it might be 80).

Answer 1

I have nutted out a solution for you.

If this is the content of input.fa:

>virusA_detailed_description
ACGTGACGATCCCCAAACGTGACGATCCCCAA
>virusB_detailed_description
ACGTGACGATCGGGAAACGTGACGATCGGGAA
>virusC_detailed_description
ACGTGACGATCAAAAAACGTGACGATCAAAAA
>virusD_detailed_description
ACGTGACGATTTTTAAACGTGACGATTTTTAA

And this is the script:

#!/usr/bin/env python3

'''
Converts a fasta file to phylip format
'''

from Bio import AlignIO
from Bio.AlignIO.PhylipIO import PhylipWriter
import sys

infile = sys.argv[1]
outfile = sys.argv[2]
idwidth = int(sys.argv[3])
with open(infile, 'r') as input_handle:
    align = AlignIO.read(input_handle, "fasta")
    with open(outfile, 'w') as output_handle:
        PhylipWriter(output_handle).write_alignment(align, id_width=idwidth)

Then, save this script as phylpwrtr.py and do:

chmod +x phylpwrtr.py
./phylpwrtr.py input.fa output_longids.phy 30 # 30 is the width of the header, adjust the value to adjust header length

The contents of output_longids.phy looks like:

 4 32
virusA_detailed_description    ACGTGACGAT CCCCAAACGT GACGATCCCC AA
virusB_detailed_description    ACGTGACGAT CGGGAAACGT GACGATCGGG AA
virusC_detailed_description    ACGTGACGAT CAAAAAACGT GACGATCAAA AA
virusD_detailed_description    ACGTGACGAT TTTTAAACGT GACGATTTTT AA