spliced_bam2gff - a tool to convert spliced BAM alignments into GFF2 format

The spliced_bam2gff tool converts BAM alignments, produced by spliced aligners (such as minimap2, gmap), into a GFF2 format.

By default, introns are created based on the N cigar feature. Alternatively, if -d (i.e. for deletion) is specified, any deletions larger than the limit will be classified as an intron. The orientation of the GFF2 features is determined by the XS strand tag and SAM flags depending on the aligner.

The tool supports splitting the output into loci and bundles of loci with a minimum number of features, which enables easy parallelisation of downstream analyses. The generated GFF2 files can be compared to a reference annotation using the gffcompare tool.

Installation

The best way to install the tool is from bioconda:

Make sure you have miniconda3 installed.
Set up the bioconda channel according to the instructions.
Install the tool by issuing conda install spliced_bam2gff

Usage

Usage of ./spliced_bam2gff:
  -L string
        Write output partitioned into "loci" to this directory. Turns of output to stdout.
  -M    Input is from minimap2.
  -S    Do NOT discard secondary and supplementary alignments.
  -V    Print out version.
  -b int
        Bundle together loci in batches with at least this size. (default -1)
  -d int
        Classify all deletions larger than this as introns (-1 means off). (default -1)
  -g    Use strand tag as feature orientation then read strand if not available.
  -h    Print out help message.
  -s    Use read strand (from BAM flag) as feature orientation.
  -t int
        Number of cores to use. (default 8)

Examples

Convert alignments generated by `minmap2 -x splice`

spliced_bam2gff -M ./test_data/sirv_simulated.bam > ./test_data/sirv_simulated_mm2.gff

Convert while classifying all deletions larger than 20 as introns:

spliced_bam2gff -d 20 -M ./test_data/sirv_simulated.bam > ./test_data/sirv_simulated_mm2.gff

Convert to GFF2 and split the output into loci separated by regions with no coverage:

spliced_bam2gff -M -L ./test_data/sirv_simulated_mm2_loci -s ./test_data/sirv_simulated.bam

Convert to GFF2 and split the output into bundles of loci with at least two thousand features:

spliced_bam2gff -b 2000 -M -L ./test_data/sirv_simulated_mm2_loci -s ./test_data/sirv_simulated.bam

Convert alignments generated by `gmap`

spliced_bam2gff  -g ./test_data/sirv_errors_gmap.bam > ./test_data/test_out_err.gff

Running tests

For running tests the following dependencies have to be installed:

gffcompare

Which is easy to install using bioconda. Look into the Makefiles for targets testing the tools on simulated and real data.

Help

Licence and Copyright

This Source Code Form is subject to the terms of the Mozilla Public License, v. 2.0. If a copy of the MPL was not distributed with this file, You can obtain one at http://mozilla.org/MPL/2.0/.

FAQs and tips

The GFF2 files can be visualised using IGV.
The GFF2 files can be converted to GFF3 or GTF using the gffread utility.

References and Supporting Information

See the post announcing the tool at the Oxford Nanopore Technologies community here.

ljyanesm/spliced_bam2gff