A lossless genomic codec — sealed core, public evidence.
A genome is reduced to a compact amortized footprint, not stored as raw sequence. On real public genomic data the codec is lossless at 193.7× on the full 24-chromosome genome — 3.11 GB reduced to 16.4 MB. Recovery is byte-identical and SHA-256 verified — a reviewer can reproduce it on public reference data and confirm the restored sequence hash-for-hash. General-purpose compressors reach 3–4× on the same data. The encoding mechanism stays sealed. Evaluation-stage — no clinical claim.
Bounded, stated-boundary claims on public reference data. The mechanism stays sealed; the density is verifiable.
The result is stated with its context: an amortized density figure on real public genomic data, where each genome reduces to a ~10 KB footprint. Recovery is byte-identical and SHA-256 chained — a reviewer can reproduce it on public reference data and verify the restored sequence hash-for-hash. General-purpose compressors reach 3–4× on the same inputs. The mechanism stays sealed; the density does not depend on trusting it.
Evaluation-stage — no clinical claim. The headline figure is one full genome — all 24 chromosomes, 3.11 GB → 16.4 MB. Amortized multi-genome cohort delta-encoding (~10 KB/genome) is a separate, higher-ratio scope. All inputs are preserved losslessly, byte-for-byte. The core projection mechanism stays sealed pending patent.
Headline 193.7× is the full 24-chr genome (3.11 GB→16.4 MB) — lossless, realistic scope, not identical-twin ratios. Amortized multi-genome cohort density is a separate scope in eval materials.
Normative files, specs, and conformance guarantees from the official repository.
Lossless genomic compression. Sealed core, public evidence.
Verifiable performance metrics compared to industry standards.
Proprietary commercial license. Use, copying, or reverse engineering of the sealed engine core is strictly prohibited. Evaluators must sign the bilateral NDA.