Science
The algorithms behind the tools, with their sources, written once where several tools share one.
A page belongs here when the method is shared rather than owned by a single tool. Two candidates were measured and left out on 2026-10-02 because only one tool uses them: guide scoring is CRISPR guide design's alone, and phylogenetic distance is Phylogenetics's alone, so each is written on that tool's page.
Each page states the method, its parameters, the citation it comes from, and its known limits. Where the engine's implementation is a faithful port of an upstream library, the page names that library and the reference outputs each result is checked against.
A tool page's The science section summarizes and links here rather than restating it.
Written
- Restriction enzymes — the dataset four tools share, REBASE as the authority, the two properties that decide whether a digest is right, and the overhang bug that shipped for months under a green test suite.
- Graph tracks — the six numbers the sequence map can plot along a molecule: GC content and skew, the cumulative skew that finds a replication origin, local composition complexity, CpG observed/expected and its island calls, and codon markers. What each one sees that the others cannot, and how to choose a window.
- Melting temperature: which model each of the tools that report a Tm computes with, why two panels can disagree about one primer, and which number to take to the bench.
- Read quality: the Phred scale, the five trimming rules the sequencing tools use and what each answers, and why masking is not trimming.
- Codon usage: the reference table every codon score is measured against, the relative adaptiveness that CAI and its relatives share, and why the highest CAI is not the best design.
- Alignment scoring: substitution matrices and why the alphabet decides which exist, linear and affine gaps, and which matrix each alignment tool uses.