pysamstats 0.24.0

pysamstats
==========

A Python utility for calculating statistics against genome positions
based on sequence alignments from a SAM or BAM file.

* Source: https://github.com/alimanfoo/pysamstats
* Download: http://pypi.python.org/pypi/pysamstats
* Release notes: https://github.com/alimanfoo/pysamstats/releases

Installation
------------

Building pysamstats depends on
[pysam version 0.8.4 or higher](http://pysam.readthedocs.org/en/latest/).
Please **install pysam first**, before attempting to install
pysamstats, e.g.:

```
$ pip install pysam>=0.8.4
$ pip install pysamstats
```

Alternatively, clone the git repo and install:

```
$ git clone git://github.com/alimanfoo/pysamstats.git
$ cd pysamstats
$ python setup.py install
```

If you have problems installing pysam, email the
[pysam user group](https://groups.google.com/forum/#!forum/pysam-user-group).

N.B., some functions also require [numpy](http://www.numpy.org) and [pytables](http://www.pytables.org) to be installed.

Usage
-----

From the command line:

```
$ pysamstats --help
Usage: pysamstats [options] FILE

Calculate statistics against genome positions based on sequence alignments
from a SAM or BAM file and print them to stdout.

Options:
-h, --help show this help message and exit
-t TYPE, --type=TYPE Type of statistics to print, one of: alignment_binned,
baseq, baseq_ext, baseq_ext_strand, baseq_strand,
coverage, coverage_binned, coverage_ext,
coverage_ext_binned, coverage_ext_strand, coverage_gc,
coverage_strand, mapq, mapq_binned, mapq_strand, tlen,
tlen_binned, tlen_strand, variation, variation_strand.
-c CHROMOSOME, --chromosome=CHROMOSOME
Chromosome name.
-s START, --start=START
Start position (1-based).
-e END, --end=END End position (1-based).
-z, --zero-based Use zero-based coordinates (default is false, i.e.,
use one-based coords).
-u, --truncate Truncate pileup-based stats so no records are emitted
outside the specified position range.
-d, --pad Pad pileup-based stats so a record is emitted for
every position (default is only covered positions).
-D MAX_DEPTH, --max-depth=MAX_DEPTH
Maximum read depth permitted in pileup-based
statistics. The default limit is 8000.
-f FASTA, --fasta=FASTA
Reference sequence file, only required for some
statistics.
-o, --omit-header Omit header row from output.
-p N, --progress=N Report progress every N rows.
--window-size=N Size of window for binned statistics (default is 300).
--window-offset=N Window offset to use for deciding which genome
position to report binned statistics against. The
default is 150, i.e., the middle of 300bp window.
--format=FORMAT Output format, one of {tsv, csv, hdf5} (defaults to
tsv). N.B., hdf5 requires PyTables to be installed.
--output=OUTPUT Path to output file. If not provided, write to stdout.
--fields=FIELDS Comma-separated list of fields to output (defaults to
all fields).
--hdf5-group=HDF5_GROUP
Name of HDF5 group to write to (defaults to the root
group).
--hdf5-dataset=HDF5_DATASET
Name of HDF5 dataset to create (defaults to "data").
--hdf5-complib=HDF5_COMPLIB
HDF5 compression library (defaults to zlib).
--hdf5-complevel=HDF5_COMPLEVEL
HDF5 compression level (defaults to 5).
--hdf5-chunksize=HDF5_CHUNKSIZE
Size of chunks in number of bytes (defaults to 2**17).

Pileup-based statistics types (each row has statistics over reads in a pileup column):

* coverage - Number of reads aligned to each genome position
(total and properly paired).
* coverage_strand - As coverage but with forward/reverse strand counts.
* coverage_ext - Various additional coverage metrics, including
coverage for reads not properly paired (mate
unmapped, mate on other chromosome, ...).
* coverage_ext_strand - As coverage_ext but with forward/reverse strand counts.
* coverage_gc - As coverage but also includes a column for %GC.
* variation - Numbers of matches, mismatches, deletions,
insertions, etc.
* variation_strand - As variation but with forward/reverse strand counts.
* tlen - Insert size statistics.
* tlen_strand - As tlen but with statistics by forward/reverse strand.
* mapq - Mapping quality statistics.
* mapq_strand - As mapq but with statistics by forward/reverse strand.
* baseq - Base quality statistics.
* baseq_strand - As baseq but with statistics by forward/reverse strand.
* baseq_ext - Extended base quality statistics, including qualities
of bases matching and mismatching reference.
* baseq_ext_strand - As baseq_ext but with statistics by forward/reverse strand.

Binned statistics types (each row has statistics over reads aligned starting within a genome window):

* coverage_binned - As coverage but binned.
* coverage_ext_binned - As coverage_ext but binned.
* mapq_binned - Similar to mapq but binned.
* alignment_binned - Aggregated counts from cigar strings.
* tlen_binned - As tlen but binned.

Examples:

pysamstats --type coverage example.bam > example.coverage.txt
pysamstats --type coverage --chromosome Pf3D7_v3_01 --start 100000 --end 200000 example.bam > example.coverage.txt

Version: 0.24 (pysam 0.8.4)
```

From Python:

```python
import pysam
import pysamstats

mybam = pysam.AlignmentFile('/path/to/your/bamfile.bam')

# iterate over statistics, one record at a time
for rec in pysamstats.stat_coverage(mybam, chrom='Pf3D7_01_v3', start=10000, end=20000):
print rec['chrom'], rec['pos'], rec['reads_all'], rec['reads_pp']
...

```

For convenience, functions are provided for loading data directly into numpy arrays, e.g.:

```python
import pysam
import pysamstats
import matplotlib.pyplot as plt

mybam = pysam.AlignmentFile('/path/to/your/bamfile.bam')
a = pysamstats.load_coverage(mybam, chrom='Pf3D7_01_v3', start=10000, end=20000)
plt.plot(a.pos, a.reads_all)
plt.show()
```

For pileup-based statistics function, note the following:

* By default a row is emitted for all genome positions covered by reads overlapping the selected region. This means rows will be emitted for positions outside the selected region, but statistics may not be accurate as not all reads overlapping that position will have been counted. To truncate output to exactly the selected region, provide a ``truncate=True`` keyword argument.
* By default a row is only emitted for genome positions covered by at least one read. To emit a row for every genome position, provide a ``pad=True`` keyword argument.
* By default the number of reads in a pileup column is limited to 8000. To increase this limit, provide a ``max_depth=100000`` keyword argument (or whatever number is suitable for your situation).

Field definitions
-----------------

The suffix **_fwd** means the field is restricted to reads mapped to
the forward strand, and **_rev** means the field is restricted to
reads mapped to the reverse strand. E.g., **reads_fwd** means the
number of reads mapped to the forward strand.

The suffix **_pp** means the field is restricted to reads flagged as
properly paired.

* **chrom** - Chromosome name.

* **pos** - Position within chromosome. One-based by default when
using the command line, zero-based by default when using the
python API.

* **reads_all** - Number of reads aligned at the position. N.b., this
is really the total, i.e., includes reads where the mate is
unmapped or otherwise not properly paired.

* **reads_pp** - Number of reads flagged as properly paired by the
aligner.

* **reads_mate_unmapped** - Number of reads where the mate is
unmapped.

* **reads_mate_other_chr** - Number of reads where the mate is mapped
to another chromosome.

* **reads_mate_same_strand** - Number of reads where the mate is
mapped to the same strand.

* **reads_faceaway** - Number of reads where the read and its mate are
mapped facing away from each other.

* **reads_softclipped** - Number of reads where there is some
softclipping at some point in the read's alignment (not
necessarily at this position).

* **reads_duplicate** - Number of reads that are flagged as duplicate.

* **gc** - Percentage GC content in the reference at this position
(depends on window length and offset specified).

* **matches** - Number of reads where the aligned base matches the
reference.

* **mismatches** - Number of reads where the aligned base does not
match the reference (but is not a deletion).

* **deletions** - Number of reads where there is a deletion in the
alignment at this position.

* **insertions** - Number of reads where there is an insertion in the
alignment at this position.

* **A/C/T/G/N** - Number of reads where the aligned base is an A/C/T/G/N.

* **mean_tlen** - Mean value of outer distance between reads and their
mates for paired reads aligned at this position. N.B., leftmost
reads in a pair have a positive tlen, rightmost reads have a
negative tlen, so if there is no strand bias, this value should be
0.

* **rms_tlen** - Root-mean-square value of outer distance between
reads and their mates for paired reads aligned at this position.

* **std_tlen** - Standard deviation of outer distance between reads
and their mates for paired reads aligned at this position.

* **reads_mapq0** - Number of reads where mapping quality is zero.

* **rms_mapq** - Root-mean-square mapping quality for reads aligned at
this position.

* **max_mapq** - Maximum value of mapping quality for reads aligned at
this position.

* **rms_baseq** - Root-mean-square value of base qualities for bases
aligned at this position.

* **rms_baseq_matches** - Root-mean-square value of base qualities for
bases aligned at this position where the base matches the
reference.

* **rms_baseq_mismatches** - Root-mean-square value of base qualities
for bases aligned at this position where the base does not match
the reference.