Labs RSS GitHub • GPL-3.0
CRV Science Labs • bio-suite v0.2.0 (Rust / GPL-3.0)

Bioinformatics, Systematics & Emergent Genome Complexity

High-throughput multi-threaded Rust worker-agent pipelines for molecular phylogenetics, structural genomics, target discovery, and information-theoretic sequence complexity.

Lab Discoveries Journal

Empirical reports from bio-suite worker-agent pipelines ranked on a 1–5★ biological signal scale.

📡 Labs RSS Feed
Signal Scale (0★ noise excluded):
★ 1/5 Baseline / Review Note ★★ 2/5 Validation ★★★ 3/5 Comparative Synthesis ★★★★ 4/5 Emergent Signal ★★★★★ 5/5 Significant Discovery
#6 ★ 1★ Baseline / Review Note Pangenomics, Complexity & Structural Drug Discovery

Pangenome Structural-Entropy & Allosteric Cryptic Target Baseline: Graph Walk Entropy in CACNA1C Intron 30 Traces CTCF Phase-Switching to an Isoform-Selective CaV1.2 Cryptic Pocket

Running the four-stage CRV Science Labs Rust bio-suite pan-cryptic pipeline (pan-cryptic-v1) reconciles pangenome variation graph topology (CACNA1C Intron 30 H_walk = 6.1732 bits, cyclomatic number beta_1 = 7, min Ollivier-Ricci curvature = -0.5000), coalescent balancing selection (Tajima's D = +2.84), nanopore 5mC CTCF insulator disruption, and scIso-Seq splicing entropy (H_iso shift from 0.8501 to 1.8929 bits) to identify a transient 468.5 A^3 allosteric pocket (PocketMiner P = 0.91 vs. 0.08 in wild-type).

12 min read • 📄 PDF • pan-cryptic-v1 Explore Discovery →
#5 ★ 1★ Baseline / Review Note Molecular Systematics & Target Discovery

Lineage-Specific Pseudogenization & Human Knockout Target Baseline: HyPhy RELAX Selection Drift Separates Long-Read Homopolymer Errors from De-Risked Inhibitory Drug Targets

Running the CRV Science Labs Rust bio-suite pseudo-loss pipeline (pseudo-loss-v1) combines TOGA/MACSE syntenic exon projection with HyPhy RELAX codon selection modeling (omega -> 1.0, k -> 0), filtering out an unpolished ONT homopolymer frameshift artifact in ABCA4 (omega = 0.082, k = 1.021) while validating true selection relaxation across primate GULO (k = 0.017) and de-risked human knockout targets ANGPTL3 (k = 0.054) and HSD17B13 (k = 0.079).

10 min read • 📄 PDF • pseudo-loss-v1 Explore Discovery →
#4 ★ 1★ Baseline / Review Note Drug Discovery & Functional Genomics

Cryptic Biosynthetic Gene Cluster & AMP Miner Baseline: MIBiG Dereplication and Embedded Self-Resistance GyrB Duplication De-Orphan a Marine Streptomyces NRPS-PKS & sORF Peptide Pair

Executing the CRV Science Labs Rust bio-suite bgc-mine pipeline (bgc-mine-v1) on the marine actinomycete benchmark Streptomyces sp. SCSIO-64 dereplicates known erythromycin-class clusters (MIBiG Jaccard = 0.924) and prioritizes two novel natural product dossiers: BGC-NRPS-SCSIO64-02 (paired with an embedded gyrB self-resistance paralog) and BGC-RIPP-SCSIO64-03 (a 31-aa cationic alpha-helical sORF peptide with hydrophobic moment mu_H = 0.6862).

9 min read • 📄 PDF • bgc-mine-v1 Explore Discovery →
#3 ★ 1★ Baseline / Review Note Genomics & Complexity Theory

Horizontal Gene Transfer & Mobile Resistome Baseline Scan: Tetranucleotide KL-Divergence Pinpoints a 20.2 kb Acinetobacter-Derived blaNDM-1 Integrative Conjugative Element in Pseudomonas

Using the CRV Science Labs Rust bio-suite hgt-scan pipeline (hgt-scan-v1), we swept a 90 kb Pseudomonas aeruginosa chromosome benchmark in 35 overlapping 5 kb windows, isolating a 20.22 kb genomic island (GI-ICE-PAE-01, peak tetranucleotide KL-divergence = 1.9759 bits, Z = +2.224) integrated at tRNA-Gly(GCC) and carrying a mobilized blaNDM-1 metallo-beta-lactamase operon.

9 min read • 📄 PDF • hgt-scan-v1 Explore Discovery →
#2 ★ 1★ Baseline / Review Note Molecular Systematics & Complexity

Carnivora Mitochondrial COI DNA Barcoding Cluster Analysis & Information-Theoretic Codon Complexity Across 16 Reference Mitogenomes

Executing the CRV Science Labs Rust bio-suite worker-agent pipeline (coi-barcode-cluster-v1) across 16 complete 1,545 bp Carnivora Cytochrome c Oxidase Subunit I (COX1) sequences resolves 4 families and 11 MOTUs while exposing Late Pleistocene peripatric mitogenome capture in Ursus, strict 2nd-codon-position Shannon entropy clamping, and near-maximal Lempel-Ziv algorithmic complexity.

12 min read • 📄 PDF • coi-barcode-cluster-v1 Explore Discovery →
#1 ★ 1★ Baseline / Review Note Structural Genomics & Information Complexity

Human GRCh38 Chromosome 1p36.33 Unbiased Sliding-Window Scan Reveals a Z = +4.24 Non-B DNA & Z-DNA Repeat Anomaly Inside the AGRN Neuromuscular Locus

Using the CRV Science Labs Rust bio-suite worker-agent pipeline (chr1-section-a-complexity-v1), we compared gene-centric CDS complexity profiling (HES4, ISG15, AGRN) against an unbiased 2,000 bp sliding-window scan across Human GRCh38 1p36.33, uncovering Region-X01 (Z = +4.2392) inside the AGRN locus.

8 min read • 📄 PDF • chr1-section-a-complexity-v1 Explore Discovery →

Code — crvscience-labs (bio-suite)

Standalone GPL-3.0 Rust 2021 codebase engineered around a Three-Tier Worker-Agent architecture with fearless rayon data-parallelism and automated core fidelity tests.

🔗 View Repository on GitHub →
# Clone the standalone GPL-3.0 repository, run the core module fidelity suite, and build release binary
git clone https://github.com/bpwhite/crvscience-labs.git
cd crvscience-labs && cargo test && cargo build --release
# Pipelines 1 & 2 — Core Barcoding & Subtelomeric Non-B DNA Complexity
./target/release/bio-suite agent coi-barcode-cluster --out runs/coi-barcode-report.json
./target/release/bio-suite agent chr1-unbiased-scan --out runs/chr1-complexity-report.json
# Part 1 (Pipelines 3–5) — Standardized High-Yield FASTA/MAG/Biobank Discovery Scanners
./target/release/bio-suite hgt-scan --out runs/hgt-scan-report.json
./target/release/bio-suite bgc-mine --out runs/bgc-mine-report.json
./target/release/bio-suite pseudo-loss --out runs/pseudo-loss-report.json
# Part 2 (Pipeline 6) — Deep Multi-Modal Pangenome Graph & Cryptic Allosteric Pocket Engine
./target/release/bio-suite pan-cryptic --out runs/pan-cryptic-report.json
./target/release/bio-suite registry

Three-Tier Agentic Pipeline Architecture

Tier 1 • Polite Data Acquisition & Local Cache

src/fetch.rs (bio-suite fetch)

Token-bucket rate-limited NCBI Entrez / Datasets fetcher (≤ 2.5 req/sec with process lockfile coordination and exponential backoff) backed by a SHA-256 content-addressable local disk cache (./cache/) and IUPAC frame-preserving FASTA/GFF3 coordinate slicers.

Tier 3 • Deterministic Worker-Agents & Emergent Composer

src/agent.rs (bio-suite agent)

Strongly typed DAG pipeline composer that seamlessly chains Tier-1 cached datasets through Tier-2 task runners without glue scripts, automatically scanning cross-engine outputs (e.g., K2P phylogenetics × codon Shannon entropy × Non-B DNA) to isolate emergent discoveries.

Tier 2 — Four Modular Scientific Engines & GPL-3.0 Drop-In Stack

1. Molecular Systematics & Phylogenomics (bio-suite phylo)

Native Rust Algorithms: Codon-anchored progressive MSA, Kimura 2-Parameter (K2P) & Tamura-Nei (TN93) distance matrices, Ti/Tv substitution bias, Neighbor-Joining (NJ) Newick inference with rayon parallel bootstrap replicates, diagnostic synapomorphy caller, and Automatic Barcode Gap Discovery (ABGD) MOTU clustering.
GPL-3.0 Drop-Ins: OrthoFinder, BUSCO, DIAMOND, MAFFT, MUSCLE5, trimAl, IQ-TREE 2, RAxML-ng, MrBayes, RevBayes, ASTRAL-IV, Dsuite, TreePL, HyPhy.

2. Structural, Population & Functional Genomics (bio-suite genome)

Native Rust Algorithms: Multi-threaded G4Hunter G-quadruplex scanner, alternating purine-pyrimidine Z-DNA propensity scorer, CpG island observed/expected analyzer, annotation-agnostic complexity anomaly peak detector, and GFF3/RefSeq coordinate intersector.
GPL-3.0 Drop-Ins: hifiasm, Verkko, Flye, Minigraph-Cactus, PGGB, vg, DeepVariant, Sniffles2, Helixer, AUGUSTUS, EDTA, Liftoff, Minimap2, MUMmer4, SyRI, GENESPACE, Cooler, MACS3.

3. Genetic Target Identification & Drug Discovery (bio-suite target)

Native Rust Algorithms: Bayesian fine-mapping Wakefield Approximate Bayes Factor (ABF) & 95% credible-set posterior inclusion probability (PIP) engine.
GPL-3.0 Drop-Ins: SuSiE, coloc, PolyFun, TwoSampleMR, S-PrediXcan, FUSION, MAGeCK, BAGEL2, rMATS-turbo, LeafCutter, Whippet, Boltz-1 (MIT complex folding), OpenFold, ESMFold, fpocket, P2Rank, AutoDock Vina, DiffDock, GNINA, OpenMM.

4. Complexity Theory & Information Genomics (bio-suite complexity)

Native Rust Algorithms: Sliding-window k-mer Shannon block entropy (H1 &dots; H6), 256-dim tetranucleotide Kullback-Leibler divergence (DKL), 64-codon RSCU, codon positional entropy (Hpos1, Hpos2, Hpos3), GC3%, Jensen-Shannon compositional divergence (DJS), Lempel-Ziv 1976 (LZ76) Kolmogorov complexity approximation, Purine-Pyrimidine DNA walk Hurst (H) & Detrended Fluctuation Analysis (αDFA), Pangenome Graph Walk Entropy (Hwalk), and Ollivier-Ricci Graph Curvature (κ).
GPL-3.0 Drop-Ins: Jellyfish, KMC3, GeCo3, odgi, panacus, HMMER, Infernal, igraph, graph-tool.

Part 1 — Three Standardized High-Yield Discovery Pipelines (src/pipelines.rs)

Pipeline 3 • Genomics & Complexity Theory

Horizontal Gene Transfer & Mobile Resistome Scanner (bio-suite hgt-scan)

1. Compositional Anomaly: 5 kb sliding-window 256-dim tetranucleotide Kullback-Leibler divergence (DKL), 64-codon RSCU cosine distance, and GC-skew excursion.
2. Mobility Boundaries: Scans 10 kb flanks for tRNA-attB/attR direct repeats (tRNAscan-SE 2.0), phage integrases (VirSorter2 / PhiSpy), insertion sequences (ISEScan), and plasmid relaxases (MOB-suite).
3. Cargo Interrogation: Screens island ORFs against AMRFinderPlus, CARD, and VFDB to flag mobilized carbapenemase and virulence cassettes.

Pipeline 4 • Drug Discovery & Genomics

Cryptic Biosynthetic Gene Cluster & AMP Miner (bio-suite bgc-mine)

1. sORF & BGC Detection: Runs Prodigal-gv + antiSMASH 7 / GECCO for NRPS, PKS, and short-ORF RiPPs/AMPs (< 100 aa).
2. MIBiG Dereplication: Filters known scaffolds exceeding 80% domain Jaccard similarity via BiG-SCAPE.
3. Self-Resistance Hook & Folding: Uses ARTS to detect duplicated, mutated core essential targets (e.g., embedded second copy of gyrB or rpoB) alongside Eisenberg amphipathic α-helical moment (μH) and ESMFold / Macrel peptide folding.

Pipeline 5 • Systematics & Target De-Risking

Lineage-Specific Pseudogenization & Trait-Loss Detector (bio-suite pseudo-loss)

1. Syntenic Exon Projection: Projects reference coding exons via TOGA / Miniprot + MACSE to call frameshift indels and premature stops (TAG/TAA/TGA).
2. Artifact vs. True Loss: Filters long-read homopolymer errors via phylogenetic sister-clade sharing (IQ-TREE 2) and mosdepth / bcftools coverage.
3. Selection Relaxation & Biobanks: Runs HyPhy RELAX (ω → 1.0, k → 0) and cross-checks gnomAD homozygous pLoF cohorts to isolate de-risked human knockout drug targets.

Part 2 — Deep Multi-Modal Engine: Pangenome Structural-Entropy & Allosteric Cryptic Target Engine (bio-suite pan-cryptic)

Pipeline 6 • Complexity Theory + Pangenomics + 3D Chromatin + Structural Drug Discovery

Reconciles HPRC Pangenome Variation Graphs (.gfa), PacBio HiFi / ONT native 5mC methylation, single-cell long-read Iso-Seq, and 3D Markov State Model protein ensembles to discover how braided structural variants alter topological graph walk entropy (Hwalk), disrupt CTCF insulators, elevate transcript splicing entropy (Hiso), and open transient, isoform-selective allosteric drug pockets.

Pipeline 6 (bio-suite pan-cryptic) — Four-Stage Multi-Modal Architecture 
[Deep Cohort Datasets: HPRC Pangenome (.gfa) + PacBio HiFi/ONT 5mC + scIso-Seq + 3D Ensembles]
                                        │
                                        ▼
         ┌──────────────────────────────────────────────────────────────┐
         │ Stage 1: Pangenome Graph Complexity & Bubble Entropy Scan    │
         │ (odgi / vg / Graph-Walk Shannon Entropy / Ollivier-Ricci)    │
         └──────────────────────────────┬───────────────────────────────┘
                                        │ (Flags Hyper-Complex Structural Hubs)
                                        ▼
         ┌──────────────────────────────────────────────────────────────┐
         │ Stage 2: Phylogenetic Coalescent & Selection on Graph Paths  │
         │ (panacus / Ancestral Recombination Graphs / Tajima's D)      │
         └──────────────────────────────┬───────────────────────────────┘
                                        │ (Identifies Adaptive / Pathogenic Haplotypes)
                                        ▼
         ┌──────────────────────────────────────────────────────────────┐
         │ Stage 3: Epigenomic Phase-Switch & Long-Read Isoform Mapping │
         │ (Modkit 5mC / IsoQuant / LeafCutter — Splicing Entropy H_iso)│
         └──────────────────────────────┬───────────────────────────────┘
                                        │ (Extracts Disease-Specific Chimeric Isoforms)
                                        ▼
         ┌──────────────────────────────────────────────────────────────┐
         │ Stage 4: Conformational Ensemble & Cryptic Pocket Discovery  │
         │ (Boltz-1 / BioEmu / OpenMM MSM / RIN Centrality / PocketMiner│
         └──────────────────────────────────────────────────────────────┘
Summary Comparison of Pipelines 3–6 (bio-suite v0.2.0)
Pipeline CLI Complexity Tier Primary Data Inputs Core Analytical Hook Discovery Deliverable
3. bio-suite hgt-scan Low (Easy Win) Draft assemblies (.fasta) KL-divergence of 4-mer spectra + RSCU + flanking integrase/tRNA calls Mobilized AMR cassettes and horizontally transferred genomic islands
4. bio-suite bgc-mine Low (Easy Win) Bacterial/Fungal genomes, MAGs antiSMASH + MIBiG dereplication + ARTS self-resistance duplication Novel natural product clusters paired with their exact molecular target
5. bio-suite pseudo-loss Low (Easy Win) Multi-species alignments, Biobank VCFs TOGA/MACSE frameshift detection + HyPhy RELAX (omega -> 1.0) Convergent trait-loss markers and safe human knockout drug targets
6. bio-suite pan-cryptic High (Deep Data) Pangenome .gfa, ONT 5mC, scIso-Seq, MD trajectories Graph walk entropy (H_walk) -> Isoform entropy (H_iso) -> PocketMiner Isoform-specific cryptic allosteric drug pockets driven by complex SVs

Apps & Distribution Releases

Precompiled standalone binaries, GPL-3.0 source distributions, and verified JSON telemetry outputs from crvscience-labs (tracked in labs-rss.xml).

Windows x86_64 Native Binary v0.2.0 • 2.42 MB • GPL-3.0

bio-suite v0.2.0 — Standalone Windows x86_64 Executable

Zero-dependency compiled Rust binary (bio-suite.exe) containing Tier-1 polite NCBI Entrez/Datasets caching, Tier-2 Phylo/Genome/Target/Complexity task engines, and all 6 Tier-3 deterministic Worker-Agent pipelines.

SHA-256: db1af789102da342a2774ad4eb9683f7ba89141aca3ec8ec4220348b2832a7ab
Cross-Platform Rust Source Distro v0.2.0 • 64.2 KB • GPL-3.0

bio-suite v0.2.0 — Complete Rust Source & Telemetry Archive

Full GPL-3.0 Rust crate source code (src/fetch.rs, src/phylo.rs, src/genome_target.rs, src/complexity.rs, src/agent.rs, src/pipelines.rs, tests/core_fidelity.rs) plus reference JSON telemetry outputs across all 6 pipelines.

SHA-256: dddb1e62ce2ed56671d301261becd9830bae6ee26e8203a3b73595d5722b5205