skip to navigation
skip to content

gtf_to_genes 1.08

Fast GTF parser

***************************************
Overview
***************************************
    We want an extremely fast, lightweight way to access gene data stored in GTF format.

    The parsed data is held in an intuitive
        Gene
             -> transcript
             -> transcript
        with exons being stored as intervals

    Our aim is to
       * cache data in binary format, which can be
       * re-read in < 10s for even the largest genomes

    Currently initial parsing Ensembl Homo sapiens release 56 takes around 4.5 minutes.
    The binary data can be reloaded in < 10s.
    This contains *all* of the data structure in the original GTF file

    Note that we sacrifice memory usage for speed. This is seldom a problem for modern computers
    and genome sizes (There are around ~400,000 exons but there are stored as intervals / int pairs)

***************************************
A Simple example
***************************************
    ::
        gene_structures = t_parse_gtf("Mus musculus")

        #
        #   used cached data for speed
        #
        ignore_cache = False

        #
        #   get all protein coding genes only
        #
        genes_by_type = gene_structures.get_genes(gtf_file, logger, ["protein_coding"], ignore_cache = ignore_cache)

        #
        #   print out gene counts
        #
        t_parse_gtf.log_gene_types (logger, genes_by_type)

        return genes_by_type