gffsub

English | 中文

Subset and QC GFF3/GTF annotations (regions, ID/attribute selectors, gene models, longest isoform, QC).

Getting Started

git clone https://github.com/WWz33/gffsub.git
cd gffsub && make -j

# minimal GFF3
printf '%s\n' '##gff-version 3' \
  $'chr1\t.\tgene\t1\t100\t.\t+\t.\tID=GeneA;Name=GeneA' \
  $'chr1\t.\tmRNA\t1\t100\t.\t+\t.\tID=mRNA1;Parent=GeneA' \
  $'chr1\t.\texon\t1\t100\t.\t+\t.\tID=exon1;Parent=mRNA1' > data/smoke.gff3

./gffsub data/smoke.gff3 -r chr1:1-100 -f gene
./gffsub data/smoke.gff3 --id GeneA -C
./gffsub data/smoke.gff3 --qc

Usage

gffsub <input.gff3> [options]
gffsub query|window|qc <input.gff3> [options]

See ./gffsub -h for the full flag list (one line per option).

Common options

Flag	Default	Description
-r, --region	—	1-based inclusive CHR:START-END
-b, --bed	—	BED intervals (0-based half-open)
--id / --ids	—	exact feature ID(s)
--name	—	gene by common naming keys
--where KEY=VALUE	—	exact column-9 attribute
-C, --children	off	include descendants of selector hits
--parents	off	include ancestors
--model	off	full gene model
--nearest REGION	—	nearest gene on same seqid
--up / --down	0	window (bp) around --id
--strand-aware	off	window follows feature strand
-f, --feature	—	type column filter
-L, --longest	off	one transcript/gene: CDS length if any CDS, else exon
-@, --threads	1	threads for --longest
--qc	off	QC report (TSV)
-t, --format	gff3	gff3|gtf|gtf2|gtf3|bed
-o, --output	stdout	output file

Input / Output

Kind	Notes
Annotation	GFF3/GTF-style feature records
GFF/GTF / -r	1-based inclusive
BED	0-based half-open
--qc TSV	severity, code, line_idx, id, message
--summary	tsv or json rows instead of features

Primary input is annotation records (not FASTA/VCF).

License

MIT License