jordeu/gendas

Data query library highly focus to work with genomic datasets

Gendas

Gendas is a "pandas" like library focus on genomic datasets. It's a data query engine highly focus to work with genomic dataset that they can always be map to a genomic position and a sequence id.

The two main goals that differentiate Gendas from other generic data query engines are:

• Native parallelization: When we apply some data manipulations in a genomic dataset, most of the times it's possible to split the genome in segments and compute each segment in parallel. Gendas tries to hide the parallelization complexity to the end users using a declarative syntax (very similar to Pandas and other dataframe libraries).
• Native data sources: In the genomics field there are several widely use data formats (tabix, bams, fasta, fastq, bed...) Gendas can directly work and mix this data formats without reading everything in memory or inserting all the data to a different data storage system.

Installation

Install the development version using pip:

pip install git+https://github.com/jordeu/gendas

Note

Gendas is still in a very early alpha development stage. Please use it as a proof of concept and to help use debug the current implementation and to improve the interaface.

Usage

Check our example scripts and notebooks to learn how to use it. And browse the documentation at readthedocs.

License

Gendas is available to the general public subject to certain conditions described in its LICENSE.

gendas's Issues

NotImplementedError: GendasMerge.__len__ not implemented

Using python 3.6 the tutorial OncodriveFML fails with this error:

NotImplementedError: GendasMerge.__len__ not implemented

This was working in previous python versions.

A possible workaround is to use a comprehension instead of the list constructor.

Replace:
list(gd['tcga'].merge(gd['cadd'], on=['REF', 'ALT'])['cadd']['PHRED'])

by:

[p for p in gd['tcga'].merge(gd['cadd'], on=['REF', 'ALT'])['cadd']['PHRED']]