Claude Skills by HolobiomicsLab

Use when you have a count matrix (genes × samples) from HTSeq, featureCounts, or transcript abundance quantification (Salmon, kallisto), a sample metadata table with experimental design, and you need to test for differential expression while controlling false discovery rate via independent.

Use when when you have salmon quant.sf.gz output files from pseudoalignment-based transcript quantification and need to convert transcript-level abundance estimates and counts into gene-level matrices for differential expression analysis with edgeR, DESeq2, or limma-voom.

Use when after running salmon quant with the --writeMappings/-z flag to produce SAM output, or when investigating discrepancies between the number of mapped reads reported in quant.sf and the actual number of records written to the output SAM file.

Use when after quantifying the same read set with two versions of a mapping/quantification tool (e.g., C++ salmon 1.11.

Use when salmon quant is run with the --writeMappings/-z flag and you need to verify that all mapped reads appear in the SAM output file.

Use when before constructing a DESeqDataSet from any count matrix (whether from tximport, HTSeq, featureCounts, or raw counts). You have sample identifiers (run IDs, file names, or row names) and must link them to condition labels (e.

Use when you have raw or minimally processed single-cell RNA-seq expression data loaded into an AnnData object (dense, sparse, or Dask-backed array as X), and you need to apply standardized preprocessing transformations (normalization, filtering, PCA) before downstream analysis such as clustering.

Use when you encounter a published scientific article or software paper that makes claims about data processing, analysis, or results, but the reproducibility context is unclear, artifacts are scattered, or the connection between claims and outputs is not immediately evident.

Use when you have observed mapping rate or quantification disagreement (e.g., >0.1% divergence in mapping rate or Pearson r < 0.

Use when you observe discrepancies in mapping rate or per-transcript quantification between two salmon implementations, or when the default chain-pruning thresholds (orphanChainSubThresh, postMergeChainSubThresh) are leaving a substantial fraction of reads unmapped (e.

Use when when comparing mapping outputs between two selective-alignment implementations (e.g., C++ vs. Rust port) on byte-identical reference indices and observing a multi-percentage-point gap in mapping rate or read assignments.

Use when you have a raw or Seurat object-backed scRNA-seq expression matrix and need to: (1) stabilize variance across genes with SCTransform normalization, (2) extract feature loadings in reduced dimensionality space via reverse PCA to use as input for GESECA or other coregulation-based enrichment.

Use when you have preprocessed single-cell RNA-seq data (normalized and dimensionality-reduced via PCA) and need to establish cell-to-cell connectivity for trajectory inference, clustering validation, or graph-based visualization.

Use when immediately after loading raw single-cell gene expression count matrices (AnnData objects) and before identifying highly variable genes or performing dimensionality reduction.

Use when when you have a raw or minimally processed scRNA-seq dataset (e.g., a Seurat object loaded from GEO) and need to prepare it for pathway enrichment or coregulation analysis.

Use when when you have computed k-nearest neighbor indices and distances (e.

Use when your input is an AnnData object with expression matrix X as a sparse scipy matrix or Dask-backed array, and you need to apply preprocessing functions (normalization, PCA, filtering) that could trigger eager materialization.

Use when after calling squidpy.gr.spatial_neighbors or any graph-building operation that outputs sparse matrices to adata.

Use when you have computed k-nearest neighbors for spatial coordinates (e.g., via pynndescent or another NN backend) and need to store the resulting adjacency and distance information in a memory-efficient format compatible with downstream graph algorithms.

Use when you have a spatial omics dataset (AnnData object with coordinate columns like 'x', 'y', 'z') and an associated tissue image file (e.g., TIFF, PNG, or HE-stained histology), and you need to extract image-based morphological features (e.

Use when after executing a spatial statistics function (e.g., squidpy.gr.sepal) on a spatial transcriptomics dataset in AnnData format, and before using the computed rankings or enrichment scores in downstream analysis.

Use when when working with spatial transcriptomics datasets (e.g., Slide-seq v2, MERFISH) stored in AnnData format and you need to identify genes whose expression shows significant spatial patterns or enrichment within tissue regions.

Use when when you have spatial molecular data (e.g., Visium, imaging-based cytometry) stored in an AnnData object with coordinate information in .

Use when after calling squidpy.gr.spatial_neighbors on an AnnData object containing spatial coordinates in obsm.

Use when you have a spatial transcriptomics experiment (e.

Use when when you have a spatial molecular dataset (e.g., Visium, MERFISH) with categorical cell-type or feature annotations and want to test whether specific categories are preferentially located near or away from each other in tissue space, beyond what random spatial distribution would predict.

Use when you have transcript-level quantification (TPM or counts from Salmon/kallisto) and need to quantify the inclusion level of specific alternative splicing events (exon skipping, intron retention, alternative splice sites, etc.) in a form suitable for differential splicing analysis across.

Use when you have run the same pathway enrichment analysis (e.

Use when you have RNA-seq count matrices (from alignment, transcript quantification, or HTSeq-count files) and need to test for differential expression between two or more conditions while controlling for batch effects or other covariates.

Use when when a statistical method offers a parameter to trade computational cost for precision (e.

Use when when you have leiden or louvain cluster assignments in single-cell data (stored in adata.obs) and need to identify cluster-specific marker genes.

Use when you have time-ordered gene expression data (e.

Use when you have modified the Scanpy codebase (e.g., added a feature or bugfix) and need to confirm that all unit and integration tests pass before submitting a pull request, or when a CI workflow fails and you need to reproduce the failure locally to diagnose the root cause.

Use when you have raw or normalized single-cell RNA-seq expression data stored in an AnnData object (`.h5ad` format) and your analysis goal is to infer developmental or differentiation trajectories.

Use when when validating a new or reimplemented quantification tool against a reference implementation on the same dataset and index, or when investigating whether changes to seed representation, chain pruning thresholds, or other algorithmic parameters affect downstream abundance estimates.

Use when you have transcript abundance files (e.g., Salmon quant.sf.gz, kallisto abundance.h5, RSEM .isoforms.results) from a quantification tool and need to construct a gene-level count matrix for DESeq2 analysis.

Use when you have a genome annotation GTF file and need to identify all transcript-level alternative splicing events (exon skipping, intron retention, alternative splice sites, mutually exclusive exons, alternative first/last exons) before quantifying their inclusion levels (PSI) across samples or.

Use when you have transcript abundance estimates from RNA-seq quantification tools (e.

Use when you have transcript-level quantification files (quant.sf.gz or quant.gz) from salmon, sailfish, kallisto, or oarfish and need to aggregate them into gene-level or transcript-level count, abundance, and length matrices for input to DESeq2, edgeR, or limma-voom.

Use when you have transcript-level abundance and count estimates from salmon, sailfish, kallisto, or oarfish and need gene-level matrices for downstream differential analysis with edgeR, DESeq2, or limma-voom.

Use when you have two implementations of the same quantification method (or major versions) and observe a persistent disagreement in mapped-read counts, per-read alignment agreement, or abundance correlations on the same reference index and read set.

Use when you have transcript quantification output (TPM or raw counts) from a pseudo-aligner (Salmon or kallisto) and an ioe/ioi event definition file from SUPPA2's generateEvents step, and you need to calculate PSI values—the relative inclusion level of alternative splicing events—across multiple.

Use when when you have transcript-level quantification files (e.g., Salmon quant.sf.gz, kallisto abundance.h5, or RSEM .results) and need to construct a gene-level count matrix for DESeq2 differential expression testing.

Use when you have transcript-level quantification files (salmon quant.sf.gz, kallisto, or Sailfish output) and need to perform gene-level differential expression analysis.

Use when after completing PCA and k-nearest neighbor graph construction on preprocessed, log-normalized, highly-variable-gene-filtered single-cell RNA-seq data (stored in an AnnData object), compute UMAP embeddings when you need a 2-D visualization for cluster inspection, cell-type annotation, or.

Use when you have PSI matrices for two or more conditions with replicates per condition, and you need to determine which alternative splicing events show significant changes between conditions while accounting for measurement uncertainty that scales with transcript expression levels.

Use when after peak detection and MS1 feature extraction from FIA-MS, GC-MS, LC-MS(/MS), or CE-MS data, when you need to identify unknown metabolites by matching observed m/z values to a reference database and want to recover HMDB identifiers, molecular formulas, and structural annotations for.

Use when when you have binned mass spectrometry imaging peaks and want to understand which detected mass-to-charge ratios represent the same metabolite in different ionization states (parent vs. adduct form).

Use when you have in silico annotations (e.g. from GNPS, timaR, or SIRIUS) paired with experimental MS/MS spectra and need to select only the highest-confidence structural matches.

Use when when you have a large collection of reference MS/MS spectra (spectral library) and need to search unknown query spectra against it rapidly, particularly for open modification searching where the modification mass is unknown and candidate space is large.