[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7EGTGPTP23VEZ2WDYZGHBDP":3,"skill-same:01M7EGTGPTP23VEZ2WDYZGHBDP":88},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":51,"commitSha":52,"skillDirectory":53,"forksCount":14,"repositoryHtmlUrl":54,"repositoryHomepageUrl":54,"ownerHtmlUrl":55,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":56,"install":57,"resources":69},"01M7EGTGPTP23VEZ2WDYZGHBDP","bio-workflows-rnaseq-to-de",null,"Orchestrates the end-to-end bulk RNA-seq differential-expression pipeline from FASTQ to an annotated DE gene table, chaining fastp QC\u002Ftrim, Salmon (decoy-aware) or STAR+featureCounts quantification, tximport gene-level collapse, DESeq2\u002FedgeR\u002Flimma-voom testing, apeglm shrinkage, and VST-based visualization. Use when committing the reference release and gene-ID namespace once for the whole run, sequencing steps in the defensible order (tximport before DE, raw counts into the model, VST only for viz\u002Fclustering), choosing alignment-free vs align-then-count and the DE engine, setting strandedness correctly, keeping batch in the design instead of correcting-then-testing, or handing the signed ranking statistic to downstream enrichment. Hands mechanism to the component skills; not a re-teach of any single step.","Mr-Nobody-Anonymous\u002FAll-skills","Mr-Nobody-Anonymous","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F184253119?v=4","01M7AEXFJHKTYQW5R0ZCQQTWFV","anonymous","01M7C00M3GQR4QKP9GZA42GFAK",0,"MIT","https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fbiology\u002Fbio-workflows-rnaseq-to-de\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: bio-workflows-rnaseq-to-de\ndescription: Orchestrates the end-to-end bulk RNA-seq differential-expression pipeline from FASTQ to an annotated DE gene table, chaining fastp QC\u002Ftrim, Salmon (decoy-aware) or STAR+featureCounts quantification, tximport gene-level collapse, DESeq2\u002FedgeR\u002Flimma-voom testing, apeglm shrinkage, and VST-based visualization. Use when committing the reference release and gene-ID namespace once for the whole run, sequencing steps in the defensible order (tximport before DE, raw counts into the model, VST only for viz\u002Fclustering), choosing alignment-free vs align-then-count and the DE engine, setting strandedness correctly, keeping batch in the design instead of correcting-then-testing, or handing the signed ranking statistic to downstream enrichment. Hands mechanism to the component skills; not a re-teach of any single step.\ntool_type: mixed\nprimary_tool: DESeq2\nworkflow: true\ndepends_on:\n  - read-qc\u002Ffastp-workflow\n  - rna-quantification\u002Falignment-free-quant\n  - read-alignment\u002Fstar-alignment\n  - read-qc\u002Frnaseq-qc\n  - rna-quantification\u002Ftximport-workflow\n  - rna-quantification\u002Fcount-matrix-qc\n  - differential-expression\u002Fdeseq2-basics\n  - differential-expression\u002Fedger-basics\n  - differential-expression\u002Fde-results\n  - differential-expression\u002Fde-visualization\nqc_checkpoints:\n  - after_qc: \"Q30 >80%, adapter content \u003C5% (RNA has a lower quality floor than DNA)\"\n  - after_quant: \"Mapping rate >70%, >10M reads mapped, flat gene-body coverage, low rRNA\u002Fintronic\"\n  - after_import: \"tx2gene release matches the Salmon index; ID conversion loses few transcripts\"\n  - after_de: \"Dispersion trend sane, PCA separates condition not batch, no Cook's outliers\"\nsource: \"https:\u002F\u002Fgithub.com\u002FGPTomics\u002FbioSkills\"\nsource_repository: \"GPTomics\u002FbioSkills\"\nsource_path: \"workflows\u002Frnaseq-to-de\u002FSKILL.md\"\nlicense: \"MIT\"\nimported_at: \"2026-09-20\"\n---\n\n## Version Compatibility\n\nReference examples tested with: DESeq2 1.42+, tximport 1.30+, apeglm 1.24+, STAR 2.7.11+, Salmon 1.10+, Subread\u002FfeatureCounts 2.0.2+ (--countReadPairs added in 2.0.2), fastp 0.23+, ggplot2 3.5+ (kallisto 0.50+ as a Salmon alternative)\n\nBefore using code patterns, verify installed versions match. If versions differ:\n- R: `packageVersion('\u003Cpkg>')` then `?function_name` to verify parameters\n- CLI: `\u003Ctool> --version` then `\u003Ctool> --help` to confirm flags\n\nIf code throws ImportError, AttributeError, or TypeError, introspect the installed\npackage and adapt the example to match the actual API rather than retrying.\n\nNote: Salmon selective alignment is default since 1.0 (the historical `--validateMappings` is now a no-op); `DESeqDataSetFromTximport` carries the average-transcript-length offset automatically; `lfcShrink(type='apeglm')` requires `coef` to name a `resultsNames(dds)` coefficient and DROPS the `stat` column. Confirm these in-tool before quoting.\n\n# RNA-seq to Differential Expression Workflow\n\n**\"Find differentially expressed genes from my RNA-seq FASTQ files\"** -> Chain QC\u002Ftrim, decoy-aware quantification, tximport gene-level collapse, a count-based DE test, shrinkage, and visualization into one annotated DE table.\n- CLI + R: fastp -> (salmon | STAR + featureCounts) -> tximport -> DESeq2\u002FedgeR\u002Flimma-voom -> lfcShrink -> VST\u002Fvolcano\n\nThis is a workflow skill: it owns the chaining decisions and hand-offs, not the internals of any one step. Every step below cross-references the component skill that teaches its mechanism.\n\n## The governing principle\n\nA bulk RNA-seq result is decided at four seams between steps, not inside any one tool.\n\n1. **The reference RELEASE + transcriptome\u002FGTF pair is a pipeline-wide commitment made once at quantification and inherited by everything downstream.** The transcriptome FASTA that builds the Salmon index and the GTF that builds the tx2gene map (and drives featureCounts) must be the SAME Ensembl\u002FGENCODE release. Mixing an index built on release 104 with a tx2gene from release 110 silently drops renamed\u002Fremoved transcripts — no error, just missing genes. This choice also fixes the gene-ID namespace (ENSG is the safe backbone; convert to symbol\u002FEntrez only at the reporting\u002Fenrichment seam). Changing the release later forces re-quantification.\n2. **Raw counts flow forward; normalized\u002Ftransformed values are terminal.** Integer counts (or tximport count-scale output) are the ONLY valid input to DESeq2\u002FedgeR\u002Flimma-voom. TPM\u002FCPM are for within-sample ranking only; a VST\u002Frlog matrix is for PCA, clustering, heatmaps, and ML — never for re-running a count-based test. Feeding the wrong scale across a join is the single most common silent corruption.\n3. **Collapse transcript->gene through tximport, not by summing counts.** tximport carries the average-transcript-length offset that corrects for isoform-usage shifts; naively summing Salmon `NumReads` biases gene counts whenever isoform usage changes across conditions.\n4. **Batch belongs in the design, not \"corrected\" then tested.** Put known batch in the formula (`~ batch + condition`). Running `removeBatchEffect`\u002FComBat and then testing on the corrected matrix exaggerates confidence (Nygaard 2016); the corrected matrix is for visualization\u002Fclustering\u002FML input only.\n\n## Pipeline map\n\n```\nFASTQ (paired)\n  | [1] QC & trim ----------------> fastp              (read-qc\u002Ffastp-workflow)\n  v\n  | [2] Quantify -----------------> salmon (decoy-aware)   (rna-quantification\u002Falignment-free-quant)\n  v     |   OR  STAR + featureCounts (need a BAM?)         (read-alignment\u002Fstar-alignment)\n  v     ^-- commitment: reference RELEASE + tx\u002FGTF pair, gene-ID namespace\n  | [3] Import & collapse tx->gene -> tximport         (rna-quantification\u002Ftximport-workflow)\n  v     ^-- carries the length offset; NEVER sum NumReads\n  | [4] Pre-DE QC ----------------> PCA \u002F dispersion \u002F outliers  (rna-quantification\u002Fcount-matrix-qc)\n  v\n  | [5] DE test ------------------> DESeq2 | edgeR-QL | limma-voom  (differential-expression\u002Fdeseq2-basics)\n  v     ^-- RAW counts in; batch in the design, not corrected-then-tested\n  | [6] Shrink & extract ---------> lfcShrink(apeglm); pull Wald `stat` for ranking  (differential-expression\u002Fde-results)\n  v\n  | [7] Visualize ----------------> VST heatmap\u002FPCA, volcano  (differential-expression\u002Fde-visualization)\n  v\nAnnotated DE table (ENSG + symbol + biotype, log2FC, stat, pvalue, padj, baseMean)\n```\n\n## Reference, IDs, and quantification target: the made-once commitments\n\nDecided before the first `salmon quant`; everything downstream inherits them. Mechanism lives in the component skills; the reasoning below is what a reviewer expects justified.\n\n| Commitment | Options | Consequence inherited downstream |\n|------------|---------|----------------------------------|\n| Reference release | One Ensembl\u002FGENCODE release for BOTH the transcriptome FASTA (index) and the GTF (tx2gene \u002F featureCounts) | Any mismatch silently drops renamed transcripts; fixes DE row names and the pathway-DB key space |\n| Gene-ID namespace | ENSG backbone (convert to symbol\u002FEntrez only at reporting) | Symbol space is lossy (aliases, many-ENSG-one-symbol merges genes); stripping the ENSG `.version` with `\\..*` also destroys the GENCODE `_PAR_Y` tag, collapsing chrY-PAR onto chrX (rna-quantification\u002Ftximport-workflow) |\n| Quantification target | Gene-level DGE (`countsFromAbundance=\"no\"`, offset carried) vs transcript-level DTU | DTU needs a DIFFERENT import (`txOut=TRUE` + `dtuScaledTPM`); switching later is a re-import, not a filter — see workflows\u002Fsplicing-pipeline |\n| 3'-tagged vs full-length | 3'-tagged (QuantSeq\u002Fbulk-10x): `countsFromAbundance=\"no\"`, no length offset | Length-bias correction does not apply to 3'-tagged libraries |\n\n## The canonical order and why\n\nEach step assumes the previous; two reorderings silently produce wrong results.\n\n1. **QC\u002Ftrim before quantification** — adapter\u002Fquality tails corrupt pseudo-mapping and duplicate structure.\n2. **Quantify to the committed reference** — Salmon decoy-aware (genome as decoy) so intron\u002Fpseudogene reads are not misassigned to transcripts; STAR only when a genome BAM is also needed downstream.\n3. **Import via tximport** — the tx->gene collapse happens HERE, carrying the length offset (order-trap: summing `NumReads` biases genes under isoform shift).\n4. **Pre-filter low-count genes** (`rowSums(counts) >= 10`) — this is a speed\u002Fmemory step, NOT the FDR filter. Order-trap: it does not replace the baseMean independent filtering that `results()` applies at the FDR step; `filterByExpr(y, design)` (edgeR) is the design-aware version and must run once BEFORE dispersion, never after.\n5. **DESeq()** on raw counts with batch in the design — size factors, dispersion, Wald\u002FLRT.\n6. **results() then lfcShrink()** — independent filtering happens inside `results()` on baseMean; shrink LFC for effect sizes\u002Franking, but p-values stay from the unshrunken test. Order-trap: apeglm\u002Fashr objects DROP the `stat` column — pull the Wald `stat` from unshrunk `results()` if a signed ranking metric is needed for GSEA.\n7. **Visualize on VST** (heatmaps\u002FPCA); volcano uses shrunken LFC + unshrunken p.\n\n## Choosing the quantifier and the DE engine\n\nPipeline-level selection only; mechanism lives in the component skills.\n\n| Fork | Lean toward | Hand off to |\n|------|-------------|-------------|\n| Alignment-free (Salmon\u002Fkallisto) vs align-then-count (STAR+featureCounts) | Salmon for gene-level DGE (decoy-aware, GC\u002Fseq-bias correction, no BAM); STAR when a genome BAM is also needed (splicing, coverage, novel junctions, variants) | rna-quantification\u002Falignment-free-quant, read-alignment\u002Fstar-alignment |\n| DESeq2 vs edgeR-QL vs limma-voom | limma-voom when library sizes vary >3x or outliers dominate; edgeR-QL for tight finite-sample type-I control; DESeq2 for the apeglm\u002Fdownstream ecosystem (70-90% top-gene overlap on well-designed data) | differential-expression\u002Fdeseq2-basics, differential-expression\u002Fedger-basics |\n| `countsFromAbundance` | `no` (gene DGE via DESeqDataSetFromTximport) \u002F `lengthScaledTPM` (DGE when the tool can't take offsets) \u002F `dtuScaledTPM`+`txOut` (DTU) | rna-quantification\u002Ftximport-workflow |\n| Strandedness `-s` | Confirm, never assume: STAR `ReadsPerGene.out.tab` cols 3 vs 4, or RSeQC `infer_experiment.py`; dUTP\u002FTruSeq is reverse (`-s 2`) | read-qc\u002Frnaseq-qc |\n\n## Primary path: Salmon + tximport + DESeq2\n\n**Goal:** turn trimmed FASTQ into a shrunken, annotated gene-level DE table.\n\n**Approach:** build a decoy-aware index once, quantify each sample, collapse to genes via tximport (release-matched tx2gene), test raw counts with batch in the design, shrink for ranking. Full runnable script: `examples\u002Fsalmon_deseq2_workflow.R`.\n\n```bash\n# Index once: decoy-aware (genome as decoy) so intron\u002Fpseudogene reads are not misassigned\ngrep \"^>\" genome.fa | cut -d \" \" -f 1 | sed 's\u002F>\u002F\u002Fg' > decoys.txt\ncat transcriptome.fa genome.fa > gentrome.fa\nsalmon index -t gentrome.fa -d decoys.txt -i salmon_index -k 31 -p 8\n\n# Quantify (selective alignment is default since 1.0; --gcBias\u002F--seqBias correct known biases)\nsalmon quant -i salmon_index -l A -1 trimmed\u002F${s}_R1.fq.gz -2 trimmed\u002F${s}_R2.fq.gz \\\n    -o quants\u002F${s} --gcBias --seqBias -p 8\n```\n\n```r\nlibrary(tximport); library(DESeq2)\n# tx2gene MUST come from the same release as the index (else renamed transcripts drop silently)\ntxi \u003C- tximport(files, type = 'salmon', tx2gene = tx2gene, ignoreTxVersion = TRUE)\ndds \u003C- DESeqDataSetFromTximport(txi, colData = coldata, design = ~ batch + condition)  # batch in design\ndds \u003C- dds[rowSums(counts(dds)) >= 10, ]              # speed filter, NOT the FDR filter\ndds$condition \u003C- relevel(dds$condition, ref = 'control')\ndds \u003C- DESeq(dds)                                     # RAW counts in\nres \u003C- lfcShrink(dds, coef = 'condition_treated_vs_control', type = 'apeglm')  # ranking\u002Feffect size\n# For GSEA ranking, pull the Wald stat from the UNSHRUNK results (apeglm drops `stat`):\nres_stat \u003C- results(dds, name = 'condition_treated_vs_control')$stat\n```\n\n## Alternative path: STAR + featureCounts + DESeq2\n\n**Goal:** produce a genome BAM (reused by splicing\u002Fcoverage\u002Fvariant steps) alongside gene counts.\n\n**Approach:** align with `--sjdbOverhang` = readlen-1, count with the verified strandedness, then `DESeqDataSetFromMatrix`. Full script: `examples\u002Fstar_deseq2_workflow.sh`.\n\n```bash\nSTAR --runMode genomeGenerate --genomeDir star_index --genomeFastaFiles genome.fa \\\n    --sjdbGTFfile genes.gtf --sjdbOverhang 149 --runThreadN 8   # 149 for 2x150, not a blanket 100\nSTAR --genomeDir star_index --readFilesIn trimmed\u002F${s}_R1.fq.gz trimmed\u002F${s}_R2.fq.gz \\\n    --readFilesCommand zcat --outSAMtype BAM SortedByCoordinate --quantMode GeneCounts \\\n    --outFileNamePrefix aligned\u002F${s}_ --runThreadN 8\n# -s from ReadsPerGene.out.tab cols 3 vs 4 (or infer_experiment.py); -s 2 = dUTP\u002FTruSeq reverse\nfeatureCounts -T 8 -p --countReadPairs -s 2 -a genes.gtf -o counts.txt aligned\u002F*_Aligned.sortedByCoord.out.bam\n```\n\n```r\ncounts \u003C- read.table('counts.txt', header = TRUE, row.names = 1, skip = 1)[, -(1:5)]\ndds \u003C- DESeqDataSetFromMatrix(countData = counts, colData = coldata, design = ~ batch + condition)\n```\n\n## QC checkpoints between steps\n\n| After | Gate | Interpretation |\n|-------|------|----------------|\n| QC\u002Ftrim | Q30 >80%, adapter \u003C5% | RNA has a lower quality floor than DNA; sharp Q30 drop = degraded input |\n| Quant\u002Falign | Mapping >70%, >10M reads mapped; flat gene-body coverage; low rRNA%\u002Fintronic% | 3' bias = degradation\u002Foligo-dT; high intronic = pre-mRNA\u002FgDNA; high intergenic = gDNA\u002Fannotation gap — all compromise DE BEFORE it runs (read-qc\u002Frnaseq-qc) |\n| Import | tx2gene release == index release; few transcripts dropped; report the ID-conversion rate (\u003C0.85 => wrong ID type or organism) | Mismatched release silently loses renamed transcripts |\n| Pre-DE | Dispersion trend sane; PCA separates condition not batch; no Cook's outliers | PCA clustering by batch => batch dominates; add it to the design (rna-quantification\u002Fcount-matrix-qc) |\n\n## Common Errors\n\n| Symptom | Cause | Fix |\n|---------|-------|-----|\n| Many genes missing \u002F low tx conversion | Salmon index and tx2gene from different releases | Rebuild both from ONE release; pin it for the whole cohort |\n| Gene counts biased where isoforms switch | Summed Salmon `NumReads` instead of importing | Collapse via tximport (carries the length offset) |\n| \"invalid class DESeqDataSet\" or nonsense LFCs | TPM\u002FVST fed into DESeq2 | Raw counts (or `lengthScaledTPM`) only; VST is for viz\u002FML |\n| Counts collapsed \u002F library looks failed | Wrong featureCounts `-s` strandedness | Infer with `infer_experiment.py` or STAR cols 3 vs 4 before counting |\n| Suspiciously many DE genes, tiny p-values | Batch-corrected matrix fed to the test | Keep batch in the design; correct only for visualization |\n| `lfcShrink` error \u002F wrong contrast | `coef` not in `resultsNames(dds)`, or ranking off the shrunk object | Use a `resultsNames` coefficient; pull `stat` from unshrunk `results()` |\n\n## Pipeline map (hand-offs)\n\n- read-qc\u002Ffastp-workflow - adapter\u002Fquality trimming and report interpretation\n- rna-quantification\u002Falignment-free-quant - Salmon\u002Fkallisto decoy-aware quantification\n- read-alignment\u002Fstar-alignment - STAR index\u002FsjdbOverhang, 2-pass, GeneCounts strandedness\n- rna-quantification\u002Ftximport-workflow - tx->gene collapse, countsFromAbundance, tx2gene, ID-version traps\n- rna-quantification\u002Fcount-matrix-qc - pre-DE PCA, dispersion, Cook's outliers, batch checks\n- differential-expression\u002Fdeseq2-basics - the DESeq2 model, design, contrasts\n- differential-expression\u002Fde-results - extracting\u002Fannotating results and the signed ranking statistic\n- differential-expression\u002Fde-visualization - volcano\u002FMA\u002Fheatmap on the right scale\n\nThe complete runnable scripts for both paths are in this skill's examples\u002F (`salmon_deseq2_workflow.R`, `star_deseq2_workflow.sh`).\n\n## Related Skills\n\n- database-access\u002Fgeo-data - Find a GSE on GEO, detect SuperSeries, link to SRA\n- database-access\u002Fsra-data - Download paired-end FASTQ from SRA \u002F ENA \u002F STRIDES cloud\n- sequence-io\u002Ffastq-quality - Confirm the FASTQ quality encoding before trimming public or pre-2011 data\n- sequence-io\u002Fpaired-end-fastq - Keep R1\u002FR2 mates synchronized; independent per-mate filtering desyncs pairs\n- read-qc\u002Ffastp-workflow - Detailed QC options and parameters\n- read-qc\u002Frnaseq-qc - Post-alignment RNA QC: strandedness, gene-body coverage, rRNA\u002Fintronic\n- read-alignment\u002Fstar-alignment - The align path (BAM for splicing\u002Fcoverage\u002Fvariants)\n- rna-quantification\u002Falignment-free-quant - Salmon and kallisto details\n- rna-quantification\u002Ftximport-workflow - tximport options, countsFromAbundance, tx2gene creation\n- rna-quantification\u002Fcount-matrix-qc - Pre-DE QC and diagnostics\n- differential-expression\u002Fdeseq2-basics - Complete DESeq2 reference\n- differential-expression\u002Fde-results - Results extraction, annotation, ranking statistic\n- differential-expression\u002Fde-visualization - Advanced visualization options\n- alternative-splicing\u002Fisoform-switching - Transcript-level DTU when gene-level is not enough (splicing fork)\n- pathway-analysis\u002Fgo-enrichment - Next step: functional enrichment (workflows\u002Fexpression-to-pathways)\n\n## References\n\n- Soneson C, Love MI, Robinson MD (2015) Differential analyses for RNA-seq: transcript-level estimates improve gene-level inferences. *F1000Research* 4:1521. DOI 10.12688\u002Ff1000research.7563.1. (tximport; the tx->gene length-offset seam.)\n- Love MI, Huber W, Anders S (2014) Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. *Genome Biology* 15:550. DOI 10.1186\u002Fs13059-014-0550-8.\n- Patro R, Duggal G, Love MI, Irizarry RA, Kingsford C (2017) Salmon provides fast and bias-aware quantification of transcript expression. *Nature Methods* 14:417-419. DOI 10.1038\u002Fnmeth.4197.\n- Nygaard V, Rødland EA, Hovig E (2016) Methods that remove batch effects while retaining group differences may lead to exaggerated confidence in downstream analyses. *Biostatistics* 17:29-39. DOI 10.1093\u002Fbiostatistics\u002Fkxv027. (batch belongs in the design.)\n- Ewels PA, Peltzer A, Fillinger S, et al (2020) The nf-core framework for community-curated bioinformatics pipelines. *Nature Biotechnology* 38:276-278. DOI 10.1038\u002Fs41587-020-0439-x. (nf-core\u002Frnaseq: the reproducible reference orchestration.)\n","## Version Compatibility\n\nReference examples tested with: DESeq2 1.42+, tximport 1.30+, apeglm 1.24+, STAR 2.7.11+, Salmon 1.10+, Subread\u002FfeatureCounts 2.0.2+ (--countReadPairs added in 2.0.2), fastp 0.23+, ggplot2 3.5+ (kallisto 0.50+ as a Salmon alternative)\n\nBefore using code patterns, verify installed versions match. If versions differ:\n- R: `packageVersion('\u003Cpkg>')` then `?function_name` to verify parameters\n- CLI: `\u003Ctool> --version` then `\u003Ctool> --help` to confirm flags\n\nIf code throws ImportError, AttributeError, or TypeError, introspect the installed\npackage and adapt the example to match the actual API rather than retrying.\n\nNote: Salmon selective alignment is default since 1.0 (the historical `--validateMappings` is now a no-op); `DESeqDataSetFromTximport` carries the average-transcript-length offset automatically; `lfcShrink(type='apeglm')` requires `coef` to name a `resultsNames(dds)` coefficient and DROPS the `stat` column. Confirm these in-tool before quoting.\n\n# RNA-seq to Differential Expression Workflow\n\n**\"Find differentially expressed genes from my RNA-seq FASTQ files\"** -> Chain QC\u002Ftrim, decoy-aware quantification, tximport gene-level collapse, a count-based DE test, shrinkage, and visualization into one annotated DE table.\n- CLI + R: fastp -> (salmon | STAR + featureCounts) -> tximport -> DESeq2\u002FedgeR\u002Flimma-voom -> lfcShrink -> VST\u002Fvolcano\n\nThis is a workflow skill: it owns the chaining decisions and hand-offs, not the internals of any one step. Every step below cross-references the component skill that teaches its mechanism.\n\n## The governing principle\n\nA bulk RNA-seq result is decided at four seams between steps, not inside any one tool.\n\n1. **The reference RELEASE + transcriptome\u002FGTF pair is a pipeline-wide commitment made once at quantification and inherited by everything downstream.** The transcriptome FASTA that builds the Salmon index and the GTF that builds the tx2gene map (and drives featureCounts) must be the SAME Ensembl\u002FGENCODE release. Mixing an index built on release 104 with a tx2gene from release 110 silently drops renamed\u002Fremoved transcripts — no error, just missing genes. This choice also fixes the gene-ID namespace (ENSG is the safe backbone; convert to symbol\u002FEntrez only at the reporting\u002Fenrichment seam). Changing the release later forces re-quantification.\n2. **Raw counts flow forward; normalized\u002Ftransformed values are terminal.** Integer counts (or tximport count-scale output) are the ONLY valid input to DESeq2\u002FedgeR\u002Flimma-voom. TPM\u002FCPM are for within-sample ranking only; a VST\u002Frlog matrix is for PCA, clustering, heatmaps, and ML — never for re-running a count-based test. Feeding the wrong scale across a join is the single most common silent corruption.\n3. **Collapse transcript->gene through tximport, not by summing counts.** tximport carries the average-transcript-length offset that corrects for isoform-usage shifts; naively summing Salmon `NumReads` biases gene counts whenever isoform usage changes across conditions.\n4. **Batch belongs in the design, not \"corrected\" then tested.** Put known batch in the formula (`~ batch + condition`). Running `removeBatchEffect`\u002FComBat and then testing on the corrected matrix exaggerates confidence (Nygaard 2016); the corrected matrix is for visualization\u002Fclustering\u002FML input only.\n\n## Pipeline map\n\n```\nFASTQ (paired)\n  | [1] QC & trim ----------------> fastp              (read-qc\u002Ffastp-workflow)\n  v\n  | [2] Quantify -----------------> salmon (decoy-aware)   (rna-quantification\u002Falignment-free-quant)\n  v     |   OR  STAR + featureCounts (need a BAM?)         (read-alignment\u002Fstar-alignment)\n  v     ^-- commitment: reference RELEASE + tx\u002FGTF pair, gene-ID namespace\n  | [3] Import & collapse tx->gene -> tximport         (rna-quantification\u002Ftximport-workflow)\n  v     ^-- carries the length offset; NEVER sum NumReads\n  | [4] Pre-DE QC ----------------> PCA \u002F dispersion \u002F outliers  (rna-quantification\u002Fcount-matrix-qc)\n  v\n  | [5] DE test ------------------> DESeq2 | edgeR-QL | limma-voom  (differential-expression\u002Fdeseq2-basics)\n  v     ^-- RAW counts in; batch in the design, not corrected-then-tested\n  | [6] Shrink & extract ---------> lfcShrink(apeglm); pull Wald `stat` for ranking  (differential-expression\u002Fde-results)\n  v\n  | [7] Visualize ----------------> VST heatmap\u002FPCA, volcano  (differential-expression\u002Fde-visualization)\n  v\nAnnotated DE table (ENSG + symbol + biotype, log2FC, stat, pvalue, padj, baseMean)\n```\n\n## Reference, IDs, and quantification target: the made-once commitments\n\nDecided before the first `salmon quant`; everything downstream inherits them. Mechanism lives in the component skills; the reasoning below is what a reviewer expects justified.\n\n| Commitment | Options | Consequence inherited downstream |\n|------------|---------|----------------------------------|\n| Reference release | One Ensembl\u002FGENCODE release for BOTH the transcriptome FASTA (index) and the GTF (tx2gene \u002F featureCounts) | Any mismatch silently drops renamed transcripts; fixes DE row names and the pathway-DB key space |\n| Gene-ID namespace | ENSG backbone (convert to symbol\u002FEntrez only at reporting) | Symbol space is lossy (aliases, many-ENSG-one-symbol merges genes); stripping the ENSG `.version` with `\\..*` also destroys the GENCODE `_PAR_Y` tag, collapsing chrY-PAR onto chrX (rna-quantification\u002Ftximport-workflow) |\n| Quantification target | Gene-level DGE (`countsFromAbundance=\"no\"`, offset carried) vs transcript-level DTU | DTU needs a DIFFERENT import (`txOut=TRUE` + `dtuScaledTPM`); switching later is a re-import, not a filter — see workflows\u002Fsplicing-pipeline |\n| 3'-tagged vs full-length | 3'-tagged (QuantSeq\u002Fbulk-10x): `countsFromAbundance=\"no\"`, no length offset | Length-bias correction does not apply to 3'-tagged libraries |\n\n## The canonical order and why\n\nEach step assumes the previous; two reorderings silently produce wrong results.\n\n1. **QC\u002Ftrim before quantification** — adapter\u002Fquality tails corrupt pseudo-mapping and duplicate structure.\n2. **Quantify to the committed reference** — Salmon decoy-aware (genome as decoy) so intron\u002Fpseudogene reads are not misassigned to transcripts; STAR only when a genome BAM is also needed downstream.\n3. **Import via tximport** — the tx->gene collapse happens HERE, carrying the length offset (order-trap: summing `NumReads` biases genes under isoform shift).\n4. **Pre-filter low-count genes** (`rowSums(counts) >= 10`) — this is a speed\u002Fmemory step, NOT the FDR filter. Order-trap: it does not replace the baseMean independent filtering that `results()` applies at the FDR step; `filterByExpr(y, design)` (edgeR) is the design-aware version and must run once BEFORE dispersion, never after.\n5. **DESeq()** on raw counts with batch in the design — size factors, dispersion, Wald\u002FLRT.\n6. **results() then lfcShrink()** — independent filtering happens inside `results()` on baseMean; shrink LFC for effect sizes\u002Franking, but p-values stay from the unshrunken test. Order-trap: apeglm\u002Fashr objects DROP the `stat` column — pull the Wald `stat` from unshrunk `results()` if a signed ranking metric is needed for GSEA.\n7. **Visualize on VST** (heatmaps\u002FPCA); volcano uses shrunken LFC + unshrunken p.\n\n## Choosing the quantifier and the DE engine\n\nPipeline-level selection only; mechanism lives in the component skills.\n\n| Fork | Lean toward | Hand off to |\n|------|-------------|-------------|\n| Alignment-free (Salmon\u002Fkallisto) vs align-then-count (STAR+featureCounts) | Salmon for gene-level DGE (decoy-aware, GC\u002Fseq-bias correction, no BAM); STAR when a genome BAM is also needed (splicing, coverage, novel junctions, variants) | rna-quantification\u002Falignment-free-quant, read-alignment\u002Fstar-alignment |\n| DESeq2 vs edgeR-QL vs limma-voom | limma-voom when library sizes vary >3x or outliers dominate; edgeR-QL for tight finite-sample type-I control; DESeq2 for the apeglm\u002Fdownstream ecosystem (70-90% top-gene overlap on well-designed data) | differential-expression\u002Fdeseq2-basics, differential-expression\u002Fedger-basics |\n| `countsFromAbundance` | `no` (gene DGE via DESeqDataSetFromTximport) \u002F `lengthScaledTPM` (DGE when the tool can't take offsets) \u002F `dtuScaledTPM`+`txOut` (DTU) | rna-quantification\u002Ftximport-workflow |\n| Strandedness `-s` | Confirm, never assume: STAR `ReadsPerGene.out.tab` cols 3 vs 4, or RSeQC `infer_experiment.py`; dUTP\u002FTruSeq is reverse (`-s 2`) | read-qc\u002Frnaseq-qc |\n\n## Primary path: Salmon + tximport + DESeq2\n\n**Goal:** turn trimmed FASTQ into a shrunken, annotated gene-level DE table.\n\n**Approach:** build a decoy-aware index once, quantify each sample, collapse to genes via tximport (release-matched tx2gene), test raw counts with batch in the design, shrink for ranking. Full runnable script: `examples\u002Fsalmon_deseq2_workflow.R`.\n\n```bash\n# Index once: decoy-aware (genome as decoy) so intron\u002Fpseudogene reads are not misassigned\ngrep \"^>\" genome.fa | cut -d \" \" -f 1 | sed 's\u002F>\u002F\u002Fg' > decoys.txt\ncat transcriptome.fa genome.fa > gentrome.fa\nsalmon index -t gentrome.fa -d decoys.txt -i salmon_index -k 31 -p 8\n\n# Quantify (selective alignment is default since 1.0; --gcBias\u002F--seqBias correct known biases)\nsalmon quant -i salmon_index -l A -1 trimmed\u002F${s}_R1.fq.gz -2 trimmed\u002F${s}_R2.fq.gz \\\n    -o quants\u002F${s} --gcBias --seqBias -p 8\n```\n\n```r\nlibrary(tximport); library(DESeq2)\n# tx2gene MUST come from the same release as the index (else renamed transcripts drop silently)\ntxi \u003C- tximport(files, type = 'salmon', tx2gene = tx2gene, ignoreTxVersion = TRUE)\ndds \u003C- DESeqDataSetFromTximport(txi, colData = coldata, design = ~ batch + condition)  # batch in design\ndds \u003C- dds[rowSums(counts(dds)) >= 10, ]              # speed filter, NOT the FDR filter\ndds$condition \u003C- relevel(dds$condition, ref = 'control')\ndds \u003C- DESeq(dds)                                     # RAW counts in\nres \u003C- lfcShrink(dds, coef = 'condition_treated_vs_control', type = 'apeglm')  # ranking\u002Feffect size\n# For GSEA ranking, pull the Wald stat from the UNSHRUNK results (apeglm drops `stat`):\nres_stat \u003C- results(dds, name = 'condition_treated_vs_control')$stat\n```\n\n## Alternative path: STAR + featureCounts + DESeq2\n\n**Goal:** produce a genome BAM (reused by splicing\u002Fcoverage\u002Fvariant steps) alongside gene counts.\n\n**Approach:** align with `--sjdbOverhang` = readlen-1, count with the verified strandedness, then `DESeqDataSetFromMatrix`. Full script: `examples\u002Fstar_deseq2_workflow.sh`.\n\n```bash\nSTAR --runMode genomeGenerate --genomeDir star_index --genomeFastaFiles genome.fa \\\n    --sjdbGTFfile genes.gtf --sjdbOverhang 149 --runThreadN 8   # 149 for 2x150, not a blanket 100\nSTAR --genomeDir star_index --readFilesIn trimmed\u002F${s}_R1.fq.gz trimmed\u002F${s}_R2.fq.gz \\\n    --readFilesCommand zcat --outSAMtype BAM SortedByCoordinate --quantMode GeneCounts \\\n    --outFileNamePrefix aligned\u002F${s}_ --runThreadN 8\n# -s from ReadsPerGene.out.tab cols 3 vs 4 (or infer_experiment.py); -s 2 = dUTP\u002FTruSeq reverse\nfeatureCounts -T 8 -p --countReadPairs -s 2 -a genes.gtf -o counts.txt aligned\u002F*_Aligned.sortedByCoord.out.bam\n```\n\n```r\ncounts \u003C- read.table('counts.txt', header = TRUE, row.names = 1, skip = 1)[, -(1:5)]\ndds \u003C- DESeqDataSetFromMatrix(countData = counts, colData = coldata, design = ~ batch + condition)\n```\n\n## QC checkpoints between steps\n\n| After | Gate | Interpretation |\n|-------|------|----------------|\n| QC\u002Ftrim | Q30 >80%, adapter \u003C5% | RNA has a lower quality floor than DNA; sharp Q30 drop = degraded input |\n| Quant\u002Falign | Mapping >70%, >10M reads mapped; flat gene-body coverage; low rRNA%\u002Fintronic% | 3' bias = degradation\u002Foligo-dT; high intronic = pre-mRNA\u002FgDNA; high intergenic = gDNA\u002Fannotation gap — all compromise DE BEFORE it runs (read-qc\u002Frnaseq-qc) |\n| Import | tx2gene release == index release; few transcripts dropped; report the ID-conversion rate (\u003C0.85 => wrong ID type or organism) | Mismatched release silently loses renamed transcripts |\n| Pre-DE | Dispersion trend sane; PCA separates condition not batch; no Cook's outliers | PCA clustering by batch => batch dominates; add it to the design (rna-quantification\u002Fcount-matrix-qc) |\n\n## Common Errors\n\n| Symptom | Cause | Fix |\n|---------|-------|-----|\n| Many genes missing \u002F low tx conversion | Salmon index and tx2gene from different releases | Rebuild both from ONE release; pin it for the whole cohort |\n| Gene counts biased where isoforms switch | Summed Salmon `NumReads` instead of importing | Collapse via tximport (carries the length offset) |\n| \"invalid class DESeqDataSet\" or nonsense LFCs | TPM\u002FVST fed into DESeq2 | Raw counts (or `lengthScaledTPM`) only; VST is for viz\u002FML |\n| Counts collapsed \u002F library looks failed | Wrong featureCounts `-s` strandedness | Infer with `infer_experiment.py` or STAR cols 3 vs 4 before counting |\n| Suspiciously many DE genes, tiny p-values | Batch-corrected matrix fed to the test | Keep batch in the design; correct only for visualization |\n| `lfcShrink` error \u002F wrong contrast | `coef` not in `resultsNames(dds)`, or ranking off the shrunk object | Use a `resultsNames` coefficient; pull `stat` from unshrunk `results()` |\n\n## Pipeline map (hand-offs)\n\n- read-qc\u002Ffastp-workflow - adapter\u002Fquality trimming and report interpretation\n- rna-quantification\u002Falignment-free-quant - Salmon\u002Fkallisto decoy-aware quantification\n- read-alignment\u002Fstar-alignment - STAR index\u002FsjdbOverhang, 2-pass, GeneCounts strandedness\n- rna-quantification\u002Ftximport-workflow - tx->gene collapse, countsFromAbundance, tx2gene, ID-version traps\n- rna-quantification\u002Fcount-matrix-qc - pre-DE PCA, dispersion, Cook's outliers, batch checks\n- differential-expression\u002Fdeseq2-basics - the DESeq2 model, design, contrasts\n- differential-expression\u002Fde-results - extracting\u002Fannotating results and the signed ranking statistic\n- differential-expression\u002Fde-visualization - volcano\u002FMA\u002Fheatmap on the right scale\n\nThe complete runnable scripts for both paths are in this skill's examples\u002F (`salmon_deseq2_workflow.R`, `star_deseq2_workflow.sh`).\n\n## Related Skills\n\n- database-access\u002Fgeo-data - Find a GSE on GEO, detect SuperSeries, link to SRA\n- database-access\u002Fsra-data - Download paired-end FASTQ from SRA \u002F ENA \u002F STRIDES cloud\n- sequence-io\u002Ffastq-quality - Confirm the FASTQ quality encoding before trimming public or pre-2011 data\n- sequence-io\u002Fpaired-end-fastq - Keep R1\u002FR2 mates synchronized; independent per-mate filtering desyncs pairs\n- read-qc\u002Ffastp-workflow - Detailed QC options and parameters\n- read-qc\u002Frnaseq-qc - Post-alignment RNA QC: strandedness, gene-body coverage, rRNA\u002Fintronic\n- read-alignment\u002Fstar-alignment - The align path (BAM for splicing\u002Fcoverage\u002Fvariants)\n- rna-quantification\u002Falignment-free-quant - Salmon and kallisto details\n- rna-quantification\u002Ftximport-workflow - tximport options, countsFromAbundance, tx2gene creation\n- rna-quantification\u002Fcount-matrix-qc - Pre-DE QC and diagnostics\n- differential-expression\u002Fdeseq2-basics - Complete DESeq2 reference\n- differential-expression\u002Fde-results - Results extraction, annotation, ranking statistic\n- differential-expression\u002Fde-visualization - Advanced visualization options\n- alternative-splicing\u002Fisoform-switching - Transcript-level DTU when gene-level is not enough (splicing fork)\n- pathway-analysis\u002Fgo-enrichment - Next step: functional enrichment (workflows\u002Fexpression-to-pathways)\n\n## References\n\n- Soneson C, Love MI, Robinson MD (2015) Differential analyses for RNA-seq: transcript-level estimates improve gene-level inferences. *F1000Research* 4:1521. DOI 10.12688\u002Ff1000research.7563.1. (tximport; the tx->gene length-offset seam.)\n- Love MI, Huber W, Anders S (2014) Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. *Genome Biology* 15:550. DOI 10.1186\u002Fs13059-014-0550-8.\n- Patro R, Duggal G, Love MI, Irizarry RA, Kingsford C (2017) Salmon provides fast and bias-aware quantification of transcript expression. *Nature Methods* 14:417-419. DOI 10.1038\u002Fnmeth.4197.\n- Nygaard V, Rødland EA, Hovig E (2016) Methods that remove batch effects while retaining group differences may lead to exaggerated confidence in downstream analyses. *Biostatistics* 17:29-39. DOI 10.1093\u002Fbiostatistics\u002Fkxv027. (batch belongs in the design.)\n- Ewels PA, Peltzer A, Fillinger S, et al (2020) The nf-core framework for community-curated bioinformatics pipelines. *Nature Biotechnology* 38:276-278. DOI 10.1038\u002Fs41587-020-0439-x. (nf-core\u002Frnaseq: the reproducible reference orchestration.)\n",{"name":5,"source":24,"license":15,"workflow":25,"tool_type":26,"depends_on":27,"description":7,"imported_at":38,"source_path":39,"primary_tool":40,"qc_checkpoints":41,"source_repository":50},"https:\u002F\u002Fgithub.com\u002FGPTomics\u002FbioSkills",true,"mixed",[28,29,30,31,32,33,34,35,36,37],"read-qc\u002Ffastp-workflow","rna-quantification\u002Falignment-free-quant","read-alignment\u002Fstar-alignment","read-qc\u002Frnaseq-qc","rna-quantification\u002Ftximport-workflow","rna-quantification\u002Fcount-matrix-qc","differential-expression\u002Fdeseq2-basics","differential-expression\u002Fedger-basics","differential-expression\u002Fde-results","differential-expression\u002Fde-visualization","2026-09-20","workflows\u002Frnaseq-to-de\u002FSKILL.md","DESeq2",[42,44,46,48],{"after_qc":43},"Q30 >80%, adapter content \u003C5% (RNA has a lower quality floor than DNA)",{"after_quant":45},"Mapping rate >70%, >10M reads mapped, flat gene-body coverage, low rRNA\u002Fintronic",{"after_import":47},"tx2gene release matches the Salmon index; ID conversion loses few transcripts",{"after_de":49},"Dispersion trend sane, PCA separates condition not batch, no Cook's outliers","GPTomics\u002FbioSkills","main","2d1750717747097b602acf30bd09a7cbd4b4c6ea","awesome_skills\u002Fbiology\u002Fbio-workflows-rnaseq-to-de","https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills","https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous","2026-11-09 11:45:36",{"source":8,"skillName":5,"commands":58,"prompt":68,"zipAvailable":25},[59,62,65],{"tool":60,"command":61},"npx","npx skills add Mr-Nobody-Anonymous\u002FAll-skills --skill bio-workflows-rnaseq-to-de",{"tool":63,"command":64},"bunx","bunx skills add Mr-Nobody-Anonymous\u002FAll-skills --skill bio-workflows-rnaseq-to-de",{"tool":66,"command":67},"pnpm","pnpm dlx skills add Mr-Nobody-Anonymous\u002FAll-skills --skill bio-workflows-rnaseq-to-de","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fbiology\u002Fbio-workflows-rnaseq-to-de\u002FSKILL.md",[70,76,82],{"path":71,"resourceType":72,"resourceTypeDesc":73,"sizeBytes":74,"sourceHtmlUrl":75},"examples\u002Fsalmon_deseq2_workflow.R","OTHER","其他",4781,"https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fbiology\u002Fbio-workflows-rnaseq-to-de\u002Fexamples\u002Fsalmon_deseq2_workflow.R",{"path":77,"resourceType":78,"resourceTypeDesc":79,"sizeBytes":80,"sourceHtmlUrl":81},"examples\u002Fstar_deseq2_workflow.sh","SCRIPT","脚本",3632,"https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fbiology\u002Fbio-workflows-rnaseq-to-de\u002Fexamples\u002Fstar_deseq2_workflow.sh",{"path":83,"resourceType":84,"resourceTypeDesc":85,"sizeBytes":86,"sourceHtmlUrl":87},"usage-guide.md","REFERENCE","参考文档",4220,"https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fbiology\u002Fbio-workflows-rnaseq-to-de\u002Fusage-guide.md",[89,97,104,111,118,125],{"skillNo":90,"name":91,"nameZh":6,"description":92,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":93,"categories":94,"keywords":95,"publishedAt":96},"01M7FFFZPC2025JQQHPHDX154J","dei-assessment","Assess diversity, equity, and inclusion with metrics. TRIGGERS - Use when user needs help with dei-assessment related tasks.","https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fhr-people\u002Fdei-assessment\u002FSKILL.md",[],[],"2026-11-16 10:05:34",{"skillNo":98,"name":99,"nameZh":6,"description":100,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":101,"categories":102,"keywords":103,"publishedAt":96},"01M7FFFWN8H4WA7A6Y398QKB6J","dei-action-plan","Create DEI action plans with goals and initiatives. TRIGGERS - Use when user needs help with dei-action-plan related tasks.","https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fhr-people\u002Fdei-action-plan\u002FSKILL.md",[],[],{"skillNo":105,"name":106,"nameZh":6,"description":107,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":108,"categories":109,"keywords":110,"publishedAt":96},"01M7FFFSHTM0QZRVZ2NRJTJ0PY","culture-fit-assessment","Design culture add assessment processes. TRIGGERS - Use when user needs help with culture-fit-assessment related tasks.","https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fhr-people\u002Fculture-fit-assessment\u002FSKILL.md",[],[],{"skillNo":112,"name":113,"nameZh":6,"description":114,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":115,"categories":116,"keywords":117,"publishedAt":96},"01M7FFFMA12RKEQA38KHJSXXPS","cowork-to-code-bridge","Use an already-installed, independently verified cowork-to-code bridge to run narrowly approved actions on the user's own macOS, Linux, or WSL2 machine through a local file queue.","https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fhr-people\u002Fcowork-to-code-bridge\u002FSKILL.md",[],[],{"skillNo":119,"name":120,"nameZh":6,"description":121,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":122,"categories":123,"keywords":124,"publishedAt":96},"01M7FFFH1F6M8YRGW5V487GH1K","counter-offer-response","Draft counter-offer responses for retention. TRIGGERS - Use when user needs help with counter-offer-response related tasks.","https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fhr-people\u002Fcounter-offer-response\u002FSKILL.md",[],[],{"skillNo":126,"name":127,"nameZh":6,"description":128,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":129,"categories":130,"keywords":131,"publishedAt":96},"01M7FFFDGEHV8Y7J9N40HZXW52","continuous-feedback-system","Design continuous feedback systems. TRIGGERS - Use when user needs help with continuous-feedback-system related tasks.","https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fhr-people\u002Fcontinuous-feedback-system\u002FSKILL.md",[],[]]