Navigation Menu

Skip to content

lrowe/hts-python

 
 

Repository files navigation

hts-python

pythonic wrapper for htslib C-API using python cffi.

There is enough functionality for this to be useful, but it still needs a lot of work.

Build Status

A taste

>>> import os.path as op

>>> from hts import Bam
>>> bam = Bam("hts/test/small.bam") #bam stolen from pybedtools [thanks]
>>> list(bam.header.seqs)
['chr2L', 'chr2R', 'chr3L', 'chr3R', 'chr4', 'chrX']

# region query creates index if needed:
>>> a = next(bam('chr2L:9000-11000'))
>>> a
Alignment('HWUSI-NAME:2:69:512:1017#0')
>>> a.target, a.pos, a.strand
('chr2L', 9329, '-')
>>> a.qlen, a.rlen
(36, 36)
>>> a.strand
'-'
>>> a.seq
'TACAAATCTTACGTAAACACTCCAAGCATGAATTCG'
>>> a.base_q[:10]
[56, 63, 53, 62, 64, 62, 51, 44, 58, 59]

>>> a.flag, a.flag_str
(16, 'REVERSE')

>>> a.cigar
Cigar('36M')

>>> str(a)[:40]
'HWUSI-NAME:2:69:512:1017#0\t16\tchr2L\t9330'

There are also wrappers for:

  • Fai for fasta querying fasta files.
  • Tbx for tabix files (indexed bed/gff/sam, etc.).
  • fisher for fisher's exact test.

Installation

  1. Install [htslib](https://github.com/samtools/htslib.git htslib) using make install
  2. pip/easy_install python cffi.
  3. run python setup.py install (--user) from this directory.

Development

This is a work in progress that relies on the hts library. All of the wrapped functions are included in hts/hts_concat.h and then available from python as, e.g. htslib.sam_read1

When C-functions not provided by the api are needed, they are added to hts_extra.c/.h.

One can run the tests with: python -c "import hts; hts.doctests()"

Things to work on:

  1. Make properties settable in hts.bam. Currently, they are read-only properties. At very least, it will be useful to have setters for seq, base_q, qname, tname, pos, strand, flag.

  2. Wrap B/VCF stuff?

Why

Why use this when pysam exists? It's an experiment with python cffi and to provide a pythonic access to htslib.

About

pythonic wrapper for libhts

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 74.3%
  • C 25.1%
  • Shell 0.6%