Methods and Citation Template
Use this page as a starting point for methods sections, benchmark notes, and software citations. Keep public statements aligned with docs/scientific-claims.md.
DotMatch uses a literal-byte alphabet policy for known-target assignment:
N and IUPAC ambiguity symbols are ordinary symbols and are not expanded as
wildcards. This policy is reported by qdaln_alphabet_policy() and recorded in
DotMatch count, demux, and pair-count summaries as alphabet_policy.
User-facing assignment workflows default to the radius ambiguity policy:
a read is unique only when exactly one target lies inside the configured
edit-distance radius. The best policy is available as an explicit
compatibility mode.
Software Citation
If you use DotMatch, cite the software release through CITATION.cff.
Installed packages also provide dotmatch citation for a copyable citation.
Use the Zenodo concept DOI 10.5281/zenodo.20541628 for general software
citation. The DOI 10.5281/zenodo.20541629 is the version DOI for v0.1.7 and
must not be used as the v0.1.9 release DOI.
Suggested citation:
O’Toole D. DotMatch: deterministic known-target short-DNA assignment for sequencing workflows. Software release v0.1.9. https://github.com/dnncha/dotmatch
Methods Sentence
For CRISPR guide-counting workflows:
Reads were assigned to the guide library using DotMatch v0.1.9 with known-target assignment, literal-byte sequence semantics, and the radius ambiguity policy. Count matrices retained only reads for which exactly one guide lay inside the configured edit-distance radius; ambiguous and unmatched reads were excluded from target counts and retained in diagnostic summaries.
For one-edit Levenshtein rescue:
DotMatch used global Levenshtein distance <=1 over the extracted guide window, including one-base substitutions, insertions, and deletions. Assignments were retained only when exactly one target was inside the configured radius, unless an explicit best-distance compatibility policy was recorded in the run summary.
For Hamming-only guide-counter-style comparisons:
DotMatch used Hamming distance <=1 over fixed-length extracted guide sequences so that the comparison matched one-substitution/no-indel guide-counting semantics.
Reproducibility Commands
Core verification:
make test
make cli-test
make python-test
make python-package-test
make repository-ready
make citation-metadata-ready
make workflow-examples-ready
make coverage
Current CRISPR evidence gates:
make public-crispr-evidence-gate
make crispr-comparison-gate
Current inline-barcode evidence gate:
make barcode-comparison-gate
Current feature-barcode assignment evidence gate:
make feature-barcode-public-gate
Current public CRISPR guide-capture assignment evidence gate:
make perturb-seq-public-gate
Current amplicon/panel primer-start assignment evidence gate:
make amplicon-panel-public-gate
Current public tiny-BCL milestone evidence gate:
make bcl-tiny-public-gate
Current oligo/adapter fixed-window public evidence gate:
make oligo-adapter-public-gate
Blocked broader comparisons:
make bcl-comparison-gate
These gates are deliberately narrow. The barcode gate is for the SRP009896/SRR391079 exact-prefix lane. The feature-barcode gate is for the 10x TotalSeq-B fixed-window per-read assignment lane, not Cell Ranger-style UMI/cell quantification. The perturb-seq public gate is for the 10x CRISPR Guide Capture fixed-window per-read assignment lane, not guide-per-cell calls, expression quantification, or perturbation-effect analysis. The amplicon/panel public gate is for the nf-core ARTIC V3 R1 fixed-window primer-start assignment lane, not consensus generation, primer trimming, variant calling, clinical panels, or diagnostic interpretation. The tiny-BCL public gate is for the public 10x tiny-BCL classic per-cycle milestone, not production demultiplexing, CBCL/NovaSeq support, or broad BCL comparison evidence. The oligo/adapter public gate is for the fast-adapter-trimming TruSeq R1 fixed-window adapter-prefix assignment lane, not adapter trimming, primer removal, UMI grouping, read merging, or production adapter workflow evidence. Leave broader BCL comparison statements out until real-data comparator evidence is in the repository.
Evidence Boundary
Describe DotMatch v0.1.9 as a known-target short-DNA assignment engine. It is not a genome aligner, general Edlib replacement, production Illumina demultiplexer, full Perturb-seq analysis pipeline, adapter trimmer, UMI grouper, read merger, or amplicon consensus/variant-calling workflow. Current public evidence supports CRISPR guide counting, exact-prefix SRP009896/SRR391079 inline-barcode demultiplexing, per-read 10x TotalSeq-B feature-barcode assignment, per-read 10x CRISPR Guide Capture assignment, nf-core ARTIC amplicon primer-start assignment, the public 10x tiny-BCL classic per-cycle milestone, fast-adapter-trimming TruSeq adapter-prefix assignment, and exact short-DNA assignment statements for the documented workflows.