@deckyard/forensicgenetics

Forensic and statistical genetics for Jennifer: STR random match probabilities with the NRC II theta correction, mixture CPI / CPE, semi-continuous (LRmix) mixture likelihood ratios, kinship and trio paternity likelihood ratios (with Familias mutation models and silent alleles), whole-pedigree likelihoods by Elston-Stewart peeling, mtDNA / Y-STR haplotype statistics, and X-STR kinship and daughter paternity likelihood ratios. For the population samples behind a table: Weir & Cockerham F-statistics (θ / F_ST) and a genotypic linkage-disequilibrium test. For reporting: the size-bias correction, a Dirichlet posterior bound on a match probability and the ENFSI verbal scale for a likelihood ratio. It works on profiles and reference allele frequencies; the only sequence input is an mtDNA region aligned to the rCRS to name its differences. Requires Jennifer 0.26.0 or newer; no deck dependencies, and it runs on jennifer-tiny.

Install

jvc add @deckyard/forensicgenetics and jvc install, or copy the deck's src/*.j, without the *_test.j overlays, into vendor/deckyard/forensicgenetics/. A release archive is deckyard-forensicgenetics-<version>.tar.gz, with the sources under deckyard-forensicgenetics-<version>/src/. Details: Installing the deck.

Example

import "@deckyard/forensicgenetics/" as forensicgenetics;

def db as forensicgenetics.FrequencyDb init forensicgenetics.withLocus(
    forensicgenetics.db("example"), "D3S1358", 1034,
    {"15": 0.2456, "16": 0.2314, "17": 0.2007, "18": 0.3223});
def suspect as forensicgenetics.Profile init forensicgenetics.profile("S1", [
    forensicgenetics.genotype("D3S1358", "15", "16")
]);

def rmp as float init forensicgenetics.randomMatchProbability($db, $suspect,
    forensicgenetics.THETA_GENERAL);
def lr as float init forensicgenetics.matchLikelihoodRatio($db, $suspect,
    forensicgenetics.THETA_GENERAL);

Documentation

https://deckyard.github.io/deck-forensicgenetics/ - concepts, casework guides, the formulae with their sources, the API reference, and the limitations to read before quoting a number.

Not a validated forensic tool

The deck implements published estimators and is tested against them, but it has not been through the validation any jurisdiction requires of casework software. Use it for research, teaching and cross-checking, and validate independently before a number informs a real conclusion. The frequency table the examples build is synthetic.

Not for legal or court use: not for a prosecution, a defence, a paternity ruling, an immigration decision, or any other legal determination. Those require validated, accredited software operated under an accredited process.

No warranty, no liability. Provided as is, without warranty of any kind, express or implied; you use it at your own risk. The author accepts no liability for any claim, damage, loss or other consequence arising from this software or any number it produces. Checking that a result is correct, appropriate and lawful for your purpose is yours alone. This restates, and does not narrow or replace, the disclaimers in the licence; where they differ, the licence governs.

Full text: Not a validated forensic tool.

Licence

LGPL-3.0-only, see LICENSE. Copyright (C) 2026 mplx <[email protected]>.