libGTF a library for parsing/searching GTF and BED files

Note that this library is not yet appropriate for real-world use (the documentation hasn't even been written yet!).

In essense, this library construct an interval tree representation of GTF or BED files and allows overlap searches, similar to GenomicRanges or bedtools. The primary difference here is the convenient C interface that allows incorporation into other programs easily.

Note that murmur3.c and murmur3.h are C implementations of MurmurHash. The C implementation is from Peter Scott and MurmurHash itself is by Austin Appleby. Both of these are in the public domain.

Installation

git clone https://github.com/dpryan79/libGTF.git
git submodule init
git submodule update
make

Examples

There are currently a few example programs in the tests/ directory.

• testBED demonstrates how to parse a BED file and display its tree representation in dot format.
• testGTF is the equivalent program for GTF files.
• testFindOverlaps demonstrates how to find overlaps of alignments in SAM/BAM/CRAM format with a GTF file. This is largely similar to featureCounts and htseq-count, though this program doesn't handle paired-end alignments intelligently (it is, afterall, just a demonstration). The code demonstrates processing sets of overlaps and merging them for further processing. This program also demonstrates how to use different match and strand types. Note that in a real program, it'd be simpler to use the findOverlapsBAM() function.
• testCountOverlaps demonstrates how to count the number of overlapping elements per alignment. Note that there's no method to filter this, for example, to only count the number of exonic overlaps. For such purposes, it's better to use the findOverlaps() function and filter the results.
• testOverlapsAny is similar to testCountOverlaps, but simply returns a binary 0 or 1, indicating whether there's at least a single overlap.
• The aforementioned programs also demonstrate how one can use filter functions, both when parsing a GTF file and when traversing one.

To Do

• compare with featureCounts/bedtools and ensure the results are the same