PyReference

A Python library for working with reference gene annotations. For RefSeq/Ensembl GRCh37/GRCh38 and other species

A GTF/GFF3 can take minutes to load. We pre-process it into JSON, so it can be loaded extremely rapidly.

PyReference makes it easy to write genomics code, which is easily run across different genomes or annotation versions.

Example

import numpy as np
from pyreference import Reference 

reference = Reference()  # uses ~/pyreference.cfg default_build

my_gene_symbols = ["MSN", "GATA2", "ZEB1"]
for gene in reference[my_gene_symbols]:
    average_length = np.mean([t.length for t in gene.transcripts])
    print("%s average length = %.2f" % (gene, average_length))
    print(gene.iv)
    for transcript in gene.transcripts:
        if transcript.is_coding:
            threep_utr = transcript.get_3putr_sequence()
            print("%s end of 3putr: %s" % (transcript.get_id(), threep_utr[-20:]))

Outputs:

MSN (MSN) 1 transcripts average length = 3970.00
chrX:[64887510,64961793)/+
NM_002444 end of 3putr: TAAAATTTAGGAAGACTTCA

GATA2 (GATA2) 3 transcripts average length = 3367.67
chr3:[128198264,128212030)/-
NM_001145662 end of 3putr: AATACTTTTTGTGAATGCCC
NM_001145661 end of 3putr: AATACTTTTTGTGAATGCCC
NM_032638 end of 3putr: AATACTTTTTGTGAATGCCC

ZEB1 (ZEB1) 6 transcripts average length = 6037.83
chr10:[31608100,31818742)/+
NM_001174093 end of 3putr: CTTCTTTTTCTATTGCCTTA
NM_001174094 end of 3putr: CTTCTTTTTCTATTGCCTTA
NM_030751 end of 3putr: CTTCTTTTTCTATTGCCTTA
NM_001174096 end of 3putr: CTTCTTTTTCTATTGCCTTA
NM_001174095 end of 3putr: CTTCTTTTTCTATTGCCTTA
NM_001128128 end of 3putr: CTTCTTTTTCTATTGCCTTA

This takes 4 seconds to load on my machine.

pyreference biotype

Also included is a command line tool (pyreference_biotype.py) which shows which biotypes small RNA fragments map to.

Installation

sudo pip install pyreference

Then you will need to:

• Download / Create gene annotations
• Create a pyreference config files