Claude Skills by GPTomics

Build the spatial neighbor graph that every downstream spatial statistic (Moran's I, neighborhood enrichment, co-occurrence, spatial domains) inherits, using Squidpy. Use when choosing the graph type (kNN vs Delaunay vs fixed-radius vs Visium hex grid) and understanding why it silently changes every downstream result; handling variable cell density (kNN fixes neighbor COUNT, fixed-radius fixes physical DISTANCE -- each distorts the other); getting coordinate units right (pixels vs microns; Vi...

Quality control, filtering, and normalization for spatial transcriptomics (Visium, Visium HD, Xenium, MERFISH/MERSCOPE, CosMx, Slide-seq) with Squidpy and Scanpy. Use when setting QC floors that do NOT delete real low-count imaging cells (an scRNA min_counts=500 floor deletes nearly every Xenium cell, whose vector is tens-to-low-hundreds of transcripts); deciding whether to normalize at all when library size carries spatial biology rather than pure technical depth; choosing cell-volume/area n...

Analyzes multiplexed antibody-imaging data (CODEX/PhenoCycler, MIBI-TOF, IMC, CyCIF, Opal/Vectra mIF) as continuous protein intensity rather than transcript counts, using scimap and squidpy. Use when choosing an intensity transform/normalization (arcsinh cofactor vs z-score vs percentile -- NOT log1p-of-counts) and correcting channel spillover and antibody-batch effects; deciding whether to phenotype by gating or by clustering on intensities; recognizing that a bounded antibody panel makes ma...

Detects spatially variable genes, spatial autocorrelation, and cell-type colocalization for spatial transcriptomics using Squidpy with PySAL/esda for local statistics. Use when choosing an SVG method by its null and scaling (SpatialDE/SPARK GP variance-component vs SPARK-X/nnSVG linear vs Moran/Geary graph autocorrelation); separating genes that are spatially variable because of cell-type composition from genes regulated within a cell type; choosing the right autocorrelation statistic (global...

Plots spatial transcriptomics expression, clusters, and annotations on tissue using Squidpy and Scanpy. Use when choosing the plotter and spot size by platform fork (sc.pl.spatial / sq.pl.spatial_scatter with real scalefactors and capture diameter for spot/capture data like Visium and Slide-seq, versus molecule/segmentation overlays for imaging/FOV data like Xenium, MERFISH, and CosMx); getting the histology coordinate-frame transform right (micron<->pixel, scalefactors) so points land on the...

Retrieves and interprets AlphaFold Protein Structure Database (AFDB) models by UniProt accession, reading pLDDT and PAE confidence correctly. Use when treating pLDDT as PER-RESIDUE confidence (not global accuracy) and recognizing a long low-pLDDT stretch as an intrinsically disordered region rather than a modeling error; reading PAE to segment confident domains and judge inter-domain/relative-position confidence that high mean pLDDT cannot certify; recognizing a static AFDB model carries NO l...

Detects putative ligand-binding pockets and druggable cavities de novo on an apo protein structure with fpocket, P2Rank, CASTp, and DoGSiteScorer, ranking them by druggability/ligandability score. Use when detecting cavities on an apo structure with no bound ligand; choosing geometric pocket enumeration (fpocket alpha-spheres, CASTp) vs ML ligandability scoring (P2Rank, DoGSiteScorer); recognizing that a geometric cavity is a hypothesis, not automatically a functional or druggable site (may b...

Measures geometric properties of protein structures with Biopython Bio.PDB - interatomic distances, distance matrices, bond and dihedral angles (phi/psi/chi, Ramachandran), superposition and RMSD, center of mass, radius of gyration, and solvent accessible surface area (SASA). Use when deciding that RMSD depends on BOTH the superposition and the atom selection (a global all-atom RMSD is dominated by flexible loops and hinge motion and is NOT a cross-protein similarity metric); choosing the met...

Maps protein-protein and protein-ligand interfaces with Bio.PDB, computing contact residues and buried surface area (BSA). Use when choosing a contact cutoff and stating its rationale (heavy-atom 4-5A vs CA-CA 8A vs a SASA-based definition); deciding a contact list is not an interface and computing buried surface area (dSASA/BSA) instead; distinguishing a genuine biological interface from a crystal-packing artifact; identifying ligand-contact or epitope residues; and computing on the biologic...

Predicts protein and complex structures with deep-learning models (ESMFold, AlphaFold2/ColabFold, AlphaFold3, Chai-1, Boltz-1/2) and reconciles them with confidence metrics. Use when choosing a predictor by input and question rather than novelty (ESMFold single-chain, no-MSA, fast, metagenomic-scale vs AlphaFold3/Chai-1/Boltz for complexes, ligands, nucleic acids, ions, PTMs); recognizing that MSA depth is the dominant accuracy determinant so ESMFold trades accuracy for speed and degrades on ...

Reads, writes, downloads, and converts macromolecular structures with Biopython Bio.PDB. Use when choosing a format (mmCIF/PDBx vs legacy PDB vs BinaryCIF) for a structure that may exceed PDB's ~62-chain / 99,999-atom limits; when residue numbers do not match the paper because of auth_* vs label_* numbering (MMCIFParser defaults auth_residues=True); when metadata (resolution, method, R-free) is missing because Bio.PDB drops it and MMCIF2Dict is needed; when the deposited coordinates are the a...

Modifies protein structures in place with Biopython Bio.PDB - transforms coordinates, strips waters/heteroatoms, overloads the B-factor column, renumbers, and builds entities. Use when applying a rotation matrix and needing to know whether it is row-convention (Entity.transform, Superimposer) or column-convention (REMARK 350 / _pdbx_struct_oper_list assembly operators) so geometry is not silently mirrored; when overloading B-factors with pLDDT/conservation for coloring and needing to preserve...

Navigate the Bio.PDB SMCRA hierarchy (Structure-Model-Chain-Residue-Atom) safely, surfacing the heterogeneity it hides by default. Use when deciding how to handle altloc/DisorderedAtom conformers before a distance or RMSD, indexing residues insertion-code-safe with the full (hetflag, resseq, icode) tuple, choosing the ATOM/observed vs SEQRES/canonical vs UniProt sequence, selecting the right Model for an NMR ensemble, filtering waters/hetero/metals correctly, and reconciling auth vs label num...

Prepares a deposited or predicted structure for docking, molecular dynamics, or electrostatics by adding hydrogens, assigning protonation and tautomer states, and filling missing atoms and short loops with PDBFixer, reduce, PROPKA, and PDB2PQR. Use when adding hydrogens an X-ray model never resolved; assigning His HID/HIE/HIP tautomers, Asn/Gln/His 180-degree flips, and Cys/Lys/Asp/Glu pKa-shifted protonation at a stated pH and microenvironment rather than trusting standard pKa 7; filling mis...

Judges whether a macromolecular model (or a region of it) is reliable enough to build on, using resolution, R-free, B-factors, MolProbity geometry, and predicted-model confidence with Bio.PDB. Use when deciding if a structure or a specific region is trustworthy before docking/mechanism/measurement; reading resolution, R-work vs R-free and the R-free-minus-R-work overfitting gap; sanity-checking per-residue and mean B-factors; flagging clashscore, Ramachandran and rotamer outliers and cis non-...

Builds and simulates multi-species metabolic community models from member genome-scale models, using MICOM for abundance-weighted steady-state community FBA and cooperative tradeoff, SMETANA for cross-feeding and competition scoring, and SteadyCom/COMETS for common-growth-rate and dynamic simulation. Use when modeling a microbiome or co-culture, predicting cross-feeding and competition, abundance-weighting members from metagenomics, choosing steady-state vs dynamic community modeling, avoidin...

Builds tissue-, cell-type-, and condition-specific metabolic models by integrating transcriptomic or proteomic data into a generic genome-scale model, using extraction algorithms (GIMME, iMAT, INIT/tINIT, MADE, E-Flux, CORDA, FASTCORE) via troppo and corda in Python or the COBRA Toolbox/RAVEN in MATLAB. Use when pruning a generic model to a context, choosing an extraction method and expression threshold, mapping expression through GPR rules to reactions, deciding whether an objective is requi...

Performs flux balance analysis (FBA), flux variability analysis (FVA), parsimonious FBA (pFBA), loopless FBA, flux sampling, and production envelopes on genome-scale metabolic models with COBRApy, solving the biomass-maximization linear program under a defined medium. Use when predicting growth rate on a carbon source, computing flux ranges and alternative optima (FVA), setting exchange bounds and minimal media, distinguishing a real growth phenotype from an under-constrained model, sampling ...

Performs in-silico single and double gene deletions, condition-dependent essentiality, and synthetic-lethality screens on genome-scale metabolic models with COBRApy, evaluating gene-protein-reaction rules and comparing FBA re-optimization against MOMA/ROOM minimal-adjustment. Use when predicting essential genes, finding synthetic-lethal pairs for drug targets, choosing a growth cutoff, deciding FBA vs MOMA vs ROOM for a knockout, making essentiality medium-specific to match an experiment, or ...

Builds draft genome-scale metabolic models from an annotated genome using CarveMe (top-down carving of a BiGG universal model) or gapseq (bottom-up pathway-evidence reconstruction), then loads and sanity-checks the draft in COBRApy. Use when creating a model for an organism without one, choosing between CarveMe and gapseq, gap-filling to a target medium, understanding why a draft that grows is still only a hypothesis, handling BiGG-vs-ModelSEED namespace mismatch, or preparing a draft for cur...

Validates, gap-fills, and standardizes genome-scale metabolic models using memote for consistency and annotation scoring and COBRApy for manual curation, including mass/charge balance, energy-generating-cycle detection, dead-end resolution, GPR fixes, and SBML/SBO/MIRIAM annotation. Use when improving a draft model, gap-filling to a target medium, detecting erroneous ATP-from-nothing cycles, interpreting a memote score correctly (consistency vs biological validity), validating predictions aga...

Computes metabolic-engineering strain designs on genome-scale models with StrainDesign (OptKnock, RobustKnock, minimal cut sets, OptCouple) and cameo (heuristic knockout and FSEOF over/under-expression targets), finding gene/reaction interventions that couple product formation to growth. Use when designing knockouts to overproduce a target chemical, choosing between OptKnock and RobustKnock, growth-coupling a product so evolution maintains it, computing minimal cut sets, finding amplification...

Reconstructs B-cell clonal families, quantifies somatic hypermutation and selection, and builds antibody lineage trees with the Immcantation R suite (alakazam, shazam, scoper, dowser, tigger) on AIRR-format BCR data. Use when deriving the clonal-clustering threshold from the distToNearest bimodal valley (never a hardcoded 0.15); choosing hierarchicalClones vs spectralClones (vj vs novj) for SHM-diverged repertoires; personalizing the germline with TIGGER before mutation counting; reconstructi...

Draws TCR/BCR repertoire figures - V-J chord/circos, CDR3 spectratype, clonal-space stratification, clonal tracking across timepoints, rarefaction/extrapolation curves, overlap heatmaps, and clonotype-similarity networks - and encodes how to read them. Use when choosing between a raw Shannon bar and a rarefaction curve for a diversity comparison; deciding a depth-robust overlap metric (Morisita-Horn) vs a set metric (Jaccard) for a heatmap; setting the distance threshold that defines a clonot...

Integrates single-cell paired TCR/BCR (10x VDJ, AIRR, dandelion, BD Rhapsody) with gene expression in an AnnData/MuData object using scirpy - chain-pairing QC, clonotype definition, clonal expansion, diversity, repertoire overlap, V(D)J usage, and VDJdb specificity. Operates on the awkward-array AIRR model (adata.obsm['airr'], accessed via get.airr after pp.index_chains), not legacy per-chain obs columns. Use when deciding clonotype definition for TCR (exact CDR3-nt identity via define_clonot...

Maps TCR/BCR receptor sequences toward candidate antigen specificity and clusters repertoires by shared-specificity signal, while enforcing that a database match or a cluster label is a HYPOTHESIS, not a specificity call. Use when deciding among database annotation (VDJdb/McPAS/IEDB+TCRMatch, requiring V-gene and HLA concordance plus a confidence score) versus sequence clustering (tcrdist3 meta-clonotypes, GLIPH2, GIANA, clusTCR, which find enrichment not per-receptor labels) versus generatio...

Computes immune-repertoire diversity, clonal structure, overlap, and segment usage from TCR/BCR clonotype tables with VDJtools (immunarch as the modern R alternative). Use when deciding which diversity estimator answers a question (q=0 observed richness/chao1/chaoE, q=1 shannonWienerIndex, q=2 inverseSimpson as a Hill profile); normalizing sequencing depth before any cross-sample claim (DownSample or the resampled CalcDiversityStats table); choosing an overlap metric (depth-robust MorisitaHor...

Tests and estimates rhythmicity at a PRE-SPECIFIED period (canonically 24h) in time-series omics using cosinor regression (CosinorPy), JTK_CYCLE/ARSER/Lomb-Scargle meta-analysis (MetaCycle meta2d), and non-parametric tests for asymmetric waveforms (RAIN, DiscoRhythm); estimates phase (acrophase), amplitude, and MESOR, and controls FDR with an effect-size (rAMP) filter against over-detection. Use when testing for 24-hour or other known-period oscillations in a single condition (circadian, feed...

Compares how a rhythm CHANGES between conditions, genotypes, treatments, tissues, or ages (differential rhythmicity), classifying each feature as gain-of-rhythm, loss-of-rhythm, phase change, amplitude change, unchanged-rhythmic, or arrhythmic-in-both, and distinguishing differential EXPRESSION (condition main effect) from differential RHYTHMICITY (condition x time interaction). Uses model-based approaches that borrow strength across conditions - LimoRhyde (sin/cos interaction terms in a limm...

Discovers a periodic signal of UNKNOWN period in time-series omics data and puts a defensible significance on it, especially when sampling is IRREGULAR (dropped timepoints, pooled harvests) so FFT/Welch/JTK are invalid. Estimates the dominant period with Lomb-Scargle / generalized Lomb-Scargle (scipy, astropy), corroborates with autocorrelation, resolves transient/time-varying periodicity with the wavelet CWT (pywt), and screens genome-wide with false-alarm probabilities under BH FDR. Use whe...

Clusters temporally variable genes by expression-profile SHAPE (not significance) using Mfuzz fuzzy c-means, TCseq, DEGreport degPatterns, and tslearn DTW/soft-DTW. Use when grouping pre-selected time-course genes into shared trajectory programs (co-expression modules), choosing between soft vs hard clustering, picking k, selecting a distance metric (Euclidean/correlation/DTW), or interpreting clusters with per-cluster enrichment. Requires temporally variable genes selected FIRST (differentia...

Infers directed, time-delayed gene regulatory edges from BULK time-series expression using Granger causality (statsmodels VAR F-test), dynGENIE3 (tree ensembles regressing ODE-derived derivatives; Random Forests by default, Extra-Trees optional), and dynamic Bayesian networks (bnlearn). Use when the output is a RANKED HYPOTHESIS list for perturbation validation, not validated causal edges; deciding Granger vs dynGENIE3 vs DBN by timepoint count and linearity; sizing maxlag against the n>3*max...

Classify variant clinical significance with the ACMG/AMP germline framework and its 2018-2025 ClinGen refinements (graded PVS1 decision tree, PM2 downgraded to Supporting, PP5/BP6 retired, calibrated PP3/BP4, Bayesian points), the AMP/ASCO/CAP somatic tiers and ClinGen oncogenicity system, ClinVar star-rating and gnomAD grpmax filtering-AF interpretation. Use when deciding germline-vs-somatic framework, applying current (not flat-2015) ACMG points, checking for a gene-specific VCEP specificat...

Generate consensus FASTA sequences by applying VCF variants onto a reference with bcftools consensus, or build viral/amplicon consensus with iVar. Use when reconstructing a sample-specific reference or haplotype, deciding -H haplotype vs IUPAC vs all-ALT projection, masking no-coverage sites so a consensus does not manufacture false reference calls, or setting iVar min-depth/min-frequency policy for surveillance genomes.

Calls germline SNPs and indels with Google DeepVariant, which reframes variant calling as CNN image classification over multi-channel pileup tensors. Covers platform-specific model selection (WGS, WES, PACBIO, ONT_R104, HYBRID_PACBIO_ILLUMINA), one-shot run_deepvariant vs the three-stage make_examples/call_variants/postprocess_variants pipeline, GPU acceleration of call_variants, DeepTrio for family/trio and de-novo calling, and joint genotyping of gVCFs with GLnexus (not GenotypeGVCFs). Use ...

Filters germline and somatic variant callsets at the site and genotype level with GATK VQSR (VQSLOD, truth-sensitivity tranches), VETS/ScoreVariantAnnotations, NVScoreVariants, hard filters with per-annotation thresholds, and bcftools/cyvcf2 expressions, plus Ti/Tv-based QC. Use when deciding between VQSR, hard filtering, and ML recalibration by cohort size and platform, setting SNP vs indel thresholds, replicating the missing-annotation-passes rule so hom-alt sites survive, applying genotype...

Joint genotype a cohort of per-sample gVCFs with GATK (HaplotypeCaller -ERC GVCF -> GenomicsDBImport or CombineGVCFs -> GenotypeGVCFs) or GLnexus for DeepVariant gVCFs, producing a squared-off sample-by-site genotype matrix. Use when deciding between joint genotyping and merging single-sample callsets (never bcftools merge as absent==hom-ref), choosing GenomicsDBImport vs CombineGVCFs by cohort size and memory, solving the N+1 problem so a new sample does not force re-calling everyone, unders...

Call structural variants (>=50 bp deletions, insertions, inversions, duplications, translocations) from short- or long-read data by reconstructing four orthogonal signals (discordant pairs, split reads via the SA tag, read depth, local assembly). Covers Manta, DELLY, LUMPY/smoove, GRIDSS2, SvABA for short reads and Sniffles2, cuteSV, pbsv, dipcall/PAV for long reads, each mapped to the signals it fuses and the blind spots that follow. Use when choosing an SV caller from its signal set and fai...

Annotates VCF variants with functional consequences, population frequencies, and pathogenicity scores using bcftools annotate/csq, Ensembl VEP, SnpEff, and ANNOVAR. Use when deciding which annotation engine and version to pin, which transcript set to report on (RefSeq vs Ensembl vs MANE Select/Plus Clinical, and why VEP --pick is dangerous clinically), how to reconcile HGVS 3'-shifting with VCF left-alignment, which consequence plus NMD status governs PVS1 eligibility, which single calibrated...

Call germline SNPs and indels from a BAM/CRAM with bcftools mpileup and call, and select the right calling engine for the job. Use when generating a VCF from aligned reads, choosing between bcftools, GATK HaplotypeCaller, DeepVariant, and DRAGEN, setting ploidy for haploid/organelle/polyploid/sex-chromosome calling, or deciding whether pileup-based calling is good enough versus a local-reassembly caller for indels and difficult regions. Not for cohort joint genotyping (see variant-calling/joi...

Left-align and trim indels to parsimonious canonical form, decompose MNPs (atomize), and split multiallelic variants with bcftools norm. Use when comparing variants across callers or cohorts, preparing a VCF for database annotation or ClinVar/dbSNP matching, merging VCFs, reconciling vt-vs-bcftools representation discordance, or resolving the VCF-left-align vs HGVS-3'-rule clash.

View, query, and interpret VCF/BCF variant files with bcftools and cyvcf2. Use when inspecting variants, extracting fields with query format strings, converting VCF/BCF, or correctly reading a field -- QUAL (site) vs GQ (genotype) vs PL/GL likelihoods, AD vs DP and allele balance, GT phasing/ploidy/PS and missing-vs-hom-ref, INFO/FORMAT Number A/R/G semantics, symbolic alleles (<DEL>, <NON_REF>, spanning *) and END, or telling a raw gVCF apart from a filtered callset.

Combine, split, sort, intersect, and subset VCF/BCF files with bcftools merge, concat, isec, sort, view, and reheader. Use when merging different samples into a cohort VCF, concatenating per-chromosome or per-region call sets for the same samples, intersecting or complementing call sets from different callers, subsetting samples/regions, harmonizing sample names and ##contig headers, or recomputing AC/AN/AF after subsetting. Covers the normalize-before-combine rule, the single-sample-merge 0/...

Compute and interpret VCF quality-control metrics (Ti/Tv, het/hom, novel/known, missingness, HWE, contamination, relatedness) with bcftools stats, vcftools, plot-vcfstats, and identity tools (somalier, peddy, KING). Use when judging whether a callset is trustworthy, diagnosing a low Ti/Tv or outlier het/hom sample, deciding whether an HWE deviation is error or biology, screening a cohort for sample swaps/contamination/wrong-sex before analysis, or comparing call sets before and after filterin...

Authors portable, strongly-typed bioinformatics pipelines in the Common Workflow Language (CWL v1.2) as CommandLineTool/Workflow/ExpressionTool documents, validated with cwltool and run at scale on Toil/Arvados/Calrissian. Use when deciding CWL (portability/provenance/regulated) vs Nextflow/WDL/Snakemake; declaring secondaryFiles for indexed companions (.bai/.fai/.dict/.tbi and the caret rule); putting resources/containers under requirements (must-hold) vs hints (advisory) to avoid silent OOM...

Authors reproducible Nextflow DSL2 pipelines built on reactive dataflow, where processes communicate only through channels and execution order is not guaranteed. Use when deciding channel/dataflow (Nextflow) vs rule-based (Snakemake) authoring; wiring queue vs value channels and fixing shared-reference exhaustion with .first(); composing DSL2 modules and subworkflows with take/main/emit; selecting container/conda profiles and pinning images by digest for portability across local/SLURM/LSF/AWS...

Runs and configures curated nf-core community Nextflow pipelines (rnaseq, sarek, atacseq, methylseq, ampliseq, taxprofiler, fetchngs) reproducibly, pinning the pipeline revision with -r and selecting a container engine and institutional config via -profile. Use when deciding to adopt a community pipeline versus author one from scratch; picking a pipeline and pinning its -r revision; selecting -profile test/docker/singularity/conda plus an institutional config from nf-core/configs; building an...

Authors reproducible bioinformatics pipelines with Snakemake - rules wired by output-file pattern, wildcards and expand() for sample fan-out, checkpoints for runtime-unknown outputs, resource/retry escalation, and conda/container software deployment on HPC and cloud. Use when deciding rule-based (Snakemake) vs channel/dataflow (Nextflow) authoring; wiring rules by OUTPUT-file pattern rather than imperative order; using wildcards + expand() for sample fan-out and constraining them to stop sile...

Authors bioinformatics pipelines in WDL (Workflow Description Language) run by Cromwell or miniwdl, targeting the GATK/Broad and Terra/AnVIL/BioData Catalyst cloud ecosystem, with tasks, workflows, scatter-gather parallelism, structs, and a runtime block that sizes the cloud VM. Use when deciding to target Terra/AnVIL/GATK/WARP (chosen for the ecosystem, not the language); sizing runtime disks dynamically for a fresh-per-task cloud VM (ceil(size(f)*factor)+buffer); choosing preemptible vs on-...

Orchestrates the end-to-end bulk ATAC-seq pipeline from FASTQ to differential accessibility and TF footprints, chaining Nextera-aware fastp QC, Bowtie2 alignment, chrM removal, dedup, a single Tn5 +4/-5 shift, MACS3 peak calling, Corces fixed-width consensus, DiffBind/csaw differential accessibility, and TOBIAS footprinting. Use when committing the reference build + blacklist once, recognizing ATAC has NO input control (the shift-extend model IS the background), applying the Tn5 shift exactly...