CRISPRware is a toolkit to preprocess NGS data and identify, score, and rank guide RNAs (gRNAs) for CRISPR experiments. It supports RNA-seq, Ribo-seq, ATAC-seq, DNase-seq, ChIP-seq, and other genomic preprocessing.

v0.2: 12 in-tree on-target scorers spanning SpCas9 and Cas12a (all bit-equivalence-tested against their upstream references), plus an in-tree Rust off-target backend (crispr-ots, ~6x faster than the previous guidescan shell-out) with bundled Cas12a CFD matrices. --cas9_scorer and --cas12a_scorer accept multiple values, so one call can emit any combination of scores.

Full documentation lives at crisprware.readthedocs.io: installation, a page per module, scoring methods, NGS-guided design, the Cas12a UCSC track pipeline, and the full command reference.

On- and off-target methods built into score_guides; each links to its documentation page:

Install per the installation docs (Python env + the crispr-ots binary on your PATH).

• FASTA (required).
• BED (optional): regions of interest, to limit the gRNA search space.
• GTF/GFF (optional): gene annotation, to filter gRNAs by feature (e.g. exon, CDS).

Demonstrated on the ce11 chromosome III FASTA + NCBI GTF in tests/test_data/ce11/ (the example off-target index is limited to chrIII, not the full ce11 genome).

index_genome builds a crispr-ots off-target index (fast, exact-mismatch) into a <name>_crisprots/ directory. score_guides auto-detects the engine from each -i index, so it can also score a Guidescan2 index (built with Guidescan2's own tooling) for RNA/DNA-bulge searches.

crisprware index_genome -f tests/test_data/ce11/chrIII_sequence.fasta -p NGG -l 20

Build gene models from the NCBI GTF:

crisprware preprocess_annotation -g tests/test_data/ce11/chrIII_ce11.ncbiRefSeq.gtf \
-m metagene consensus longest shortest

Default settings generate NGG protospacer guides:

crisprware generate_guides -f tests/test_data/ce11/chrIII_sequence.fasta \
-k tests/test_data/ce11/chrIII_ce11.ncbiRefSeq.gtf \
--feature CDS

Scoring takes ~5 minutes and uses 8 threads by default (--threads <int>). --tracr is Chen2013, Hsu2013, or both (see RuleSet3). Pass more than one -i index to get a separate specificity_<index> column from each engine in one run.

crisprware score_guides -b chrIII_sequence_gRNA/chrIII_sequence_gRNA.bed \
-i chrIII_sequence_crisprots/chrIII_sequence_crisprots --tracr Chen2013 --threads 8

Rank on scoring columns: -c paired with -m filters (here RS3_score_Chen2013 >= 0 and specificity_chrIII_sequence_crisprots >= 0.2); -p 5 65 -f CDS keeps the 5th-65th CDS percentile; --output_all writes a TSV + histograms per filter stage.

crisprware rank_guides \
-k chrIII_sequence_scoredgRNA/chrIII_sequence_scoredgRNA.bed \
-t tests/test_data/ce11/chrIII_ce11.ncbiRefSeq.gtf \
-f CDS \
-c RS3_score_Chen2013 specificity_chrIII_sequence_crisprots \
-m 0 0.2 \
-p 5 65 \
-r RS3_score_Chen2013 \
--output_all