gumpy

Genetics with Numpy

Installation

git clone https://github.com/oxfordmmm/gumpy
cd gumpy
pip install .

Documentation

Easy access to documentation for public methods can be found using the pydoc module from a terminal:

python -m pydoc -b gumpy

This should open a browser window showing documentation for all loaded modules. Navigating to gumpy (package) should bring up available files to view documentation.

Docstrings contain documentation for almost all methods if documentation of private methods is required.

Testing

A suite of tests can be run from a terminal:

python -m pytest --cov=gumpy -vv

Usage

Parse a genbank file

Genome objects can be created by passing a filename of a genbank file

from gumpy import Genome

g = Genome("filename.gbk")

Parse a VCF file

VCFFile objects can be created by passing a filename of a vcf file

from gumpy import VCFFile

vcf = VCFFile("filename.vcf")

Apply a VCF file to a reference genome

The mutations defined in a vcf file can be applied to a reference genome to produce a new Genome object containing the changes detailed in the vcf.

If a contig is set within the vcf, the length of the contig should match the length of the genome. Otherwise, if the vcf details changes within the genome range, they will be made.

from gumpy import Genome, VCFFile

reference_genome = Genome("reference.gbk")
vcf = VCFFile("filename.vcf")

resultant_genome = reference_genome + vcf

Genome level comparisons

There are two different methods for comparing changes. One can quickly check for changes which are caused by a given VCF file. The other can check for changes between two genome. The latter is therefore suited best for comparisons in which either both genomes are mutated, or the VCF file(s) are not available. The former is best suited for cases where changes caused by a VCF want to be determined, but finding gene-level differences will require rebuilding the Gene objects, which can be time consuming.

Compare genomes

Two genomes of the same length can be easily compared, including equality and changes between the two. Best suited to cases where two mutated genomes are to be compared.

from gumpy import Genome, GenomeDifference

g1 = Genome("filename1.gbk")
g2 = Genome("filename2.gbk")

diff = g2 - g1 #Genome.difference returns a GenomeDifference object
print(diff.snp_distance) #SNP distance between the two genomes
print(diff.variants) #Array of variants (SNPs/INDELs) of the differences between g2 and g1

Gene level comparisons

When a Genome object is instanciated, it is populated with Gene objects for each gene detailed in the genbank file. These genes can also be compared. Gene differences can be found through direct comparison of Gene objects, or systematically through the gene_differences() method of GenomeDifference.

from gumpy import Genome, Gene

g1 = Genome("filename1.gbk")
g2 = Genome("filename2.gbk")

#Get the Gene objects for the gene "gene1_name" from both Genomes
g1_gene1 = g1.build_gene["gene1_name"]
g2_gene1 = g2.build_gene["gene1_name"]

g1_gene1 == g2_gene1 #Equality check of the two genes
diff= g1_gene1 - g2_gene1 #Returns a GeneDifference object
diff.mutations #List of mutations in GARC describing the variation between the two genes

Save and load Genome objects

Due to how long it takes to create a Genome object, it may be beneficial to save the object to disk. The reccomendation is to utilise the pickle module to do so, but due to the security implications of this, do so at your own risk! An example is below:

import pickle

import gumpy

#Load genome
g = gumpy.Genome("filename.gbk")

#Save genome
pickle.dump(g, open("filename.pkl", "wb"))

#Load genome
g2 = pickle.load(open("filename.pkl", "rb"))

g == g2 #True