flowalign 0.10

Alignment generator using mappy and Python

FlowAlign

FlowAlign is a package to simplify realignment to a reference for viral sequences, using only Python packages. (Under the hood it is using minimap2, so it's fast.)

Installation

pip install flowalign

Usage (command-line)

flowalign sequences.fasta --reference wuhCor1.fa --output aligned.fa

If you omit the output parameter, output will go to STDOUT.

Usage (Python):

First we download the reference, and some unaligned sequences to align to it:

wget https://hgdownload.soe.ucsc.edu/goldenPath/wuhCor1/bigZips/wuhCor1.fa.gz && gunzip wuhCor1.fa.gz
wget https://data.nextstrain.org/files/ncov/open/global/sequences.fasta.xz &&  xz --decompress sequences.fasta.xz

Then we write a simple Python script, creating an iterator called aligned that will yield reference-aligned versions of these unaligned sequences:

import flowalign
aligned = flowalign.yield_aligned(input="sequences.fasta", reference= "wuhCor1.fa")
for name, aligned_sequence in aligned:
    print(">"+name)
    print(aligned_sequence)

yield_aligned can also take a stream, e.g.:

aligned = flowalign.yield_aligned(input=open("sequences.fasta","rt), reference= "wuhCor1.fa")

Under the hood, mappy is in some sense calling minimap2 with --secondary=no --sam-hit-only --score-N=0 -x asm5.

Note that the multiprocessing implementation is fairly hacky which may cause issues. It is mostly expected that yield_aligned will be only called once at any one time.

Acknowledgements

This package is developed by me, but it owes almost everything to sam_2_fasta by Ben Jackson at the University of Edinburgh -- major functions are taken from that codebase. (Ben has also ported this code to gofasta). Sam2fasta is typically run on a SAM file from minimap2. flowalign incorporates the alignment process using mappy, the Python-bindings for minimap2. The idea is that one doesn't need any dependencies except Python packages (mappy supplies its own minimap2) to get aligned sequences.