[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7ECVBMEEG7AAP7DNME392PC":3,"skill-same:01M7ECVBMEEG7AAP7DNME392PC":70},{"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":30,"commitSha":31,"skillDirectory":32,"forksCount":14,"repositoryHtmlUrl":33,"repositoryHomepageUrl":33,"ownerHtmlUrl":34,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":35,"install":36,"resources":49},"01M7ECVBMEEG7AAP7DNME392PC","bio-differential-expression-de-visualization",null,"Creates DE-specific diagnostic and result visualizations using DESeq2\u002FedgeR built-in functions and lightweight ggplot2 wrappers. Covers MA plot (with the shrunken-LFC compression effect), volcano (with the apeglm caveat that p-values are unchanged), PCA on VST\u002Frlog (never raw counts), sample distance heatmaps, top-DE-gene heatmaps with the row-scaling trap, dispersion \u002F BCV plot interpretation, p-value histogram diagnostics, plotCounts for individual genes, blind=TRUE vs FALSE rationale, and the n=3 visualization stake. Use when generating DE diagnostic plots, choosing VST vs rlog for visualization, troubleshooting suspicious plot patterns (shifted MA cloud, batch-dominated PCA, anti-conservative p-value histogram), or building a standard QC figure panel.","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-differential-expression-de-visualization\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: bio-differential-expression-de-visualization\ndescription: Creates DE-specific diagnostic and result visualizations using DESeq2\u002FedgeR built-in functions and lightweight ggplot2 wrappers. Covers MA plot (with the shrunken-LFC compression effect), volcano (with the apeglm caveat that p-values are unchanged), PCA on VST\u002Frlog (never raw counts), sample distance heatmaps, top-DE-gene heatmaps with the row-scaling trap, dispersion \u002F BCV plot interpretation, p-value histogram diagnostics, plotCounts for individual genes, blind=TRUE vs FALSE rationale, and the n=3 visualization stake. Use when generating DE diagnostic plots, choosing VST vs rlog for visualization, troubleshooting suspicious plot patterns (shifted MA cloud, batch-dominated PCA, anti-conservative p-value histogram), or building a standard QC figure panel.\ntool_type: r\nprimary_tool: DESeq2\nsource: \"https:\u002F\u002Fgithub.com\u002FGPTomics\u002FbioSkills\"\nsource_repository: \"GPTomics\u002FbioSkills\"\nsource_path: \"differential-expression\u002Fde-visualization\u002FSKILL.md\"\nlicense: \"MIT\"\nimported_at: \"2026-09-20\"\n---\n\n## Version Compatibility\n\nReference examples tested with: DESeq2 1.42+, edgeR 4.0+, limma 3.58+, ggplot2 3.5+, pheatmap 1.0+, RColorBrewer 1.1+, ggrepel 0.9+, EnhancedVolcano 1.20+, matrixStats 1.2+\n\nBefore using code patterns, verify installed versions match. If versions differ:\n- R: `packageVersion('\u003Cpkg>')` then `?function_name` to verify parameters\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\n# DE Visualization\n\n**\"Make the standard DE figure panel\"** -> Use built-in functions or thin wrappers to produce diagnostic plots (dispersion, p-value histogram, PCA, sample distance) and result plots (MA, volcano, heatmap of top DE genes, per-gene counts), interpreted as diagnostics of the underlying model.\n\n## Scope\n\nThis skill covers DE-specific built-in plots and immediate wrappers. For richer customization:\n- Custom volcano\u002FMA with apeglm-shrunken LFC and ggrepel labelling -> `data-visualization\u002Fvolcano-and-ma-plots`\n- PCA \u002F UMAP \u002F t-SNE customization -> `data-visualization\u002Fdimensionality-reduction-plots`\n- Heatmap customization and ComplexHeatmap recipes -> `data-visualization\u002Fheatmaps-clustering`\n\n## The Single Most Important Modern Insight -- A volcano with shrunken LFC compresses the cloud, but the p-values are unchanged\n\n`lfcShrink()` pulls noisy estimates toward zero. On the volcano, that pulls genes horizontally toward the center. But the y-axis (`-log10(pvalue)`) is the **unshrunken Wald p-value** -- shrinkage does NOT recompute p-values (Zhu, Ibrahim, Love 2019 *Bioinformatics* 35:2084). A naive reader sees fewer extreme dots and concludes \"fewer genes are significant\". Wrong: the same genes are significant; the effect sizes are smaller and more honest.\n\nAlways label the volcano x-axis \"shrunken log2 fold change (apeglm)\" and note the y-axis comes from the unshrunken Wald test. The whole point of the apeglm volcano is the honest effect-size axis; if a publication shows an unshrunken volcano, it is showing inflated effects from low-count noise.\n\nThe MA plot has its own version of this: shrinkage flattens the left side (low-mean, formerly extreme LFC) and barely touches the right (high-mean, well-estimated LFC). That asymmetry is the visual signature of working shrinkage.\n\n## Plot Taxonomy\n\n| Plot | Diagnostic OR result | Built-in function | What it tests |\n|------|---------------------|-------------------|---------------|\n| Dispersion plot | Diagnostic | `plotDispEsts(dds)` (DESeq2), `plotBCV(y)` (edgeR) | Mean-dispersion trend fit quality |\n| p-value histogram | Diagnostic | None; use ggplot2 | Null calibration, hidden batch, over-correction |\n| PCA on VST\u002Frlog | Diagnostic + result | `plotPCA(vsd, intgroup=...)` (DESeq2), `plotMDS()` (edgeR via limma) | Sample clustering, batch effects, outliers |\n| Sample distance heatmap | Diagnostic | `pheatmap` on `dist(t(assay(vsd)))` | Within-group consistency, sample swaps |\n| MA plot | Diagnostic + result | `plotMA(res)` (DESeq2), `plotMD(qlf)` (edgeR) | Normalization sanity, LFC vs mean |\n| Volcano | Result | ggplot2 wrapper; `EnhancedVolcano` | Top-effect, top-significance gene story |\n| Top-DE heatmap | Result | `pheatmap` on `assay(vsd)[sig_genes,]` | Per-gene pattern across conditions |\n| `plotCounts` per gene | Result | `plotCounts(dds, gene, intgroup)` | Per-gene biology |\n\n## Decision Tree by Scenario\n\n| Scenario | Recommended approach |\n|----------|---------------------|\n| PCA for unbiased QC | `vst(dds, blind = TRUE)`; ask \"do samples group as expected without design influence?\" |\n| PCA for results figure | `vst(dds, blind = FALSE)`; design is settled, accept its influence on dispersion |\n| n \u003C 30, library sizes vary >4x | `rlog(dds, blind = FALSE)` instead of vst |\n| n > 30 | `vst()`; rlog impractical |\n| Volcano | Plot shrunken LFC on x, unshrunken p-value on y; label both axes |\n| Sample distance heatmap | `vst(blind = TRUE)`; tells if a sample is the wrong group regardless of design |\n| Top-DE heatmap, want to see PATTERN | `scale = 'row'` (z-score per gene) |\n| Top-DE heatmap, want to see ABSOLUTE LEVEL | `scale = 'none'` on `assay(vsd)`; otherwise weak signal looks strong |\n| Top-variable-gene selection | `matrixStats::rowMads(assay(vsd))` instead of `rowVars` -- MAD is outlier-robust |\n| n = 3, top genes in volcano | Note Schurch 2016 finding: 20-40% of true positives missed; treat as exploratory |\n| Many groups, comparing DE sets | UpSet plot (Lex 2014); Venn drowns above 3 sets |\n\n## Dispersion Diagnostic (Run This First)\n\n**Goal:** Verify the dispersion-mean trend was fit acceptably before trusting any results.\n\n**Approach:** `plotDispEsts(dds)` (DESeq2) or `plotBCV(y)` (edgeR) shows gene-wise (black\u002Fblue), fitted trend (red), and final shrunken (blue) dispersions vs mean.\n\n```r\nplotDispEsts(dds)\n\nplotBCV(y)\n```\n\n| Pattern | Meaning | Action |\n|---------|---------|--------|\n| Cloud follows trend; final shrunken estimates pulled toward red curve | Healthy fit | Proceed |\n| Red trend nowhere near the gene-wise cloud | Parametric trend failed | `DESeq(dds, fitType = 'local')` or `fitType = 'mean'` |\n| Many gene-wise dispersions FAR ABOVE the trend | Outlier or unmodeled batch genes | Inspect rather than trust QL F-test alone |\n| Final estimates much lower than gene-wise everywhere | Excessive shrinkage; sample too small or trend too flat | Check `useEM`, robust hyperparameter setting |\n| BCV decreases monotonically with mean | Correct in edgeR | Default trend |\n\nA plot inspected before trusting results is worth a hundred lines of statistical safeguards.\n\n## P-value Histogram (Run This Second)\n\n**Goal:** Detect model misspecification or hidden batch before reporting any gene list.\n\n**Approach:** Histogram of raw p-values; under a correctly specified null, uniform with a spike near zero.\n\n```r\nlibrary(ggplot2)\nggplot(res_df, aes(x = pvalue)) +\n    geom_histogram(bins = 50, fill = 'steelblue', color = 'white') +\n    labs(x = 'P-value', y = 'Frequency', title = 'P-value distribution') +\n    theme_bw()\n```\n\n| Shape | Meaning | Action |\n|-------|---------|--------|\n| Uniform + spike at 0 | Correctly specified | Proceed |\n| U-shape (spikes at 0 AND 1) | Anti-conservative; hidden batch or unmodeled covariate | Add the missing covariate; re-fit |\n| Depleted near 0, spike near 1 | Conservative; over-modeled or wrong dispersion | Simplify model; check dispersion plot |\n| Spike only at p = 1 | Discrete artifact from very-low-count genes | Pre-filter more aggressively |\n\n## MA Plot (LFC vs Mean)\n\n**Goal:** Inspect the relationship between LFC and mean expression for normalization correctness and shrinkage effect.\n\n**Approach:** `plotMA` (DESeq2) or `plotMD` (edgeR). Always pick `ylim` deliberately; default can flatten the signal.\n\n```r\nplotMA(res, ylim = c(-5, 5), main = 'MA plot (unshrunken)')\n\nres_apeglm \u003C- lfcShrink(dds, coef = 'condition_treated_vs_control', type = 'apeglm')\nplotMA(res_apeglm, ylim = c(-5, 5), main = 'MA plot (apeglm-shrunken)')\n\nplotMD(qlf, main = 'edgeR MD plot')\nabline(h = c(-1, 1), col = 'blue', lty = 2)\n```\n\n| Pattern | Meaning |\n|---------|---------|\n| Symmetric cloud centered at LFC = 0 | Correct normalization |\n| Cloud median clearly above or below 0 | Normalization failed (TMM\u002FRLE assumption violated) -- see normalization skill |\n| Funnel widening at low mean | Expected (low counts noisier) |\n| Dramatic up\u002Fdown asymmetry | Possibly real (large biological perturbation), possibly normalization failure -- cross-check |\n| Discrete horizontal bands at low mean | Low-count artifacts; pre-filter more aggressively |\n\nThe apeglm-shrunken MA visually flattens the left side; the post-shrinkage cloud should be tighter at low means.\n\n## Volcano with Shrunken LFC\n\n**Goal:** Show effect size vs significance with honest fold changes.\n\n**Approach:** Use a built-in renderer (EnhancedVolcano for quick publication-quality output) on shrunken LFCs. Always plot shrunken LFC; always set `max.overlaps = Inf` when labeling >10 genes -- the ggrepel default (10) silently drops labels. EnhancedVolcano accepts `max.overlaps` directly in 1.12+; version 1.10-1.11 has the older `maxoverlapsConnectors` argument (default 15); for either, falling back to `options(ggrepel.max.overlaps = Inf)` at the top of the script also works. For full ggplot2 customization (color schemes, faceting, label-set engineering), see `data-visualization\u002Fvolcano-and-ma-plots`.\n\n```r\nlibrary(EnhancedVolcano)\n\nres_apeglm \u003C- lfcShrink(dds, coef = 'condition_treated_vs_control', type = 'apeglm')\n\nEnhancedVolcano(res_apeglm,\n    lab = rownames(res_apeglm),\n    x = 'log2FoldChange', y = 'pvalue',\n    pCutoff = 0.05, FCcutoff = 1,\n    title = 'Treatment vs Control',\n    subtitle = 'Shrunken LFC (apeglm); unshrunken Wald p',\n    max.overlaps = Inf)\n```\n\n## PCA on VST\u002Frlog (Never on Raw Counts)\n\n**Goal:** Show sample clustering by condition; detect batch effects, swaps, outliers.\n\n**Approach:** Variance-stabilize first (VST or rlog), THEN PCA. Raw counts make PC1 = library size; log(counts+1) makes PC1 = mean expression. Neither carries biological signal until variance is stabilized.\n\n```r\nvsd \u003C- vst(dds, blind = FALSE)\nplotPCA(vsd, intgroup = c('condition', 'batch'))\n\npca_df \u003C- plotPCA(vsd, intgroup = c('condition', 'batch'), returnData = TRUE)\npercentVar \u003C- round(100 * attr(pca_df, 'percentVar'))\n\nlibrary(ggplot2)\nggplot(pca_df, aes(PC1, PC2, color = condition, shape = batch)) +\n    geom_point(size = 4) +\n    xlab(paste0('PC1: ', percentVar[1], '% variance')) +\n    ylab(paste0('PC2: ', percentVar[2], '% variance')) +\n    theme_bw()\n\nlibrary(limma)\nplotMDS(cpm(y, log = TRUE), col = as.numeric(group), pch = 16)\n```\n\n`blind=TRUE` (default for `vst()`) re-estimates dispersions ignoring the design -- appropriate for unbiased QC (\"are samples consistent independent of design?\"). `blind=FALSE` uses the fitted dispersions -- appropriate for downstream visualization where the design is settled. Modern DESeq2 vignette recommends `blind=FALSE` for any plot after the model is fit.\n\n| PCA pattern | Interpretation | Action |\n|-------------|----------------|--------|\n| Clear separation by condition on PC1 or PC2 | Strong biological signal | Proceed |\n| Separation by batch, not condition | Batch effect dominates | Include batch in design; DO NOT subtract before DE (see batch-correction Nygaard 2016) |\n| One sample far from its group | Outlier or swap | Check library QC; sex check; somalier |\n| Condition signal on PC3+, not PC1-PC2 | Subtle effect | May still find DE; review dispersion plot |\n| Two distinct sample clusters not explained by metadata | Hidden covariate | Investigate processing date, lane, machine |\n\n## Sample Distance Heatmap (for QC)\n\n```r\nlibrary(pheatmap)\nvsd \u003C- vst(dds, blind = TRUE)\nsd \u003C- dist(t(assay(vsd)))\nmat \u003C- as.matrix(sd)\nann \u003C- data.frame(condition = colData(dds)$condition,\n                  row.names = colnames(dds))\npheatmap(mat, annotation_col = ann, annotation_row = ann,\n         clustering_distance_rows = sd, clustering_distance_cols = sd,\n         color = colorRampPalette(c('white', 'steelblue'))(100),\n         main = 'Sample distance (vst blind)')\n```\n\nThe diagonal should be dark; within-group samples should cluster. A within-group sample distant from its peers is a candidate for sample swap.\n\n## Top-DE Heatmap and the Row-Scaling Trap\n\n**Goal:** Show expression patterns of significant genes across samples for results figure.\n\n**Approach:** Use `vst(blind=FALSE)`, select top genes (by adjusted p-value or MAD-robust variance), choose scaling deliberately.\n\n```r\nlibrary(pheatmap)\n\nsig \u003C- rownames(subset(res, padj \u003C 0.01))[1:50]\nvsd \u003C- vst(dds, blind = FALSE)\nmat \u003C- assay(vsd)[sig, ]\n\nmat_scaled \u003C- t(scale(t(mat)))\n\nann_col \u003C- data.frame(condition = colData(dds)$condition,\n                      batch     = colData(dds)$batch,\n                      row.names = colnames(mat))\n\npheatmap(mat_scaled, annotation_col = ann_col,\n         show_rownames = FALSE,\n         clustering_distance_rows = 'correlation',\n         clustering_distance_cols = 'correlation',\n         color = colorRampPalette(c('blue', 'white', 'red'))(100),\n         main = 'Top 50 DE genes (z-scored per gene)')\n```\n\n`scale='row'` (z-score per gene) is the conventional choice for \"show me patterns\". It DESTROYS absolute expression level information -- a gene at 5-7 with mean 6 looks identical to a gene at 10-1000. For pattern detection: correct. For QC heatmaps showing batch shifts: WRONG -- use `scale='none'` on `assay(vsd)`.\n\nTop-variable-gene selection robustness:\n\n```r\nlibrary(matrixStats)\nvars_mad \u003C- rowMads(assay(vsd))\ntop500 \u003C- order(vars_mad, decreasing = TRUE)[1:500]\n```\n\n`rowMads` (median absolute deviation) is outlier-robust; `rowVars` is dominated by single-outlier-sample genes. For exploratory PCA of \"top variable genes\", MAD selection avoids artifacts.\n\n## Per-gene Plot\n\n```r\nplotCounts(dds, gene = 'GENE_NAME', intgroup = 'condition')\n\nd \u003C- plotCounts(dds, gene = 'GENE_NAME', intgroup = c('condition','batch'),\n                returnData = TRUE)\nlibrary(ggplot2)\nggplot(d, aes(x = condition, y = count, color = batch)) +\n    geom_jitter(width = 0.1, size = 3) +\n    scale_y_log10() +\n    ggtitle('GENE_NAME') +\n    theme_bw()\n```\n\nWith n=3, the boxplot is misleading (3 points per box). Prefer `geom_jitter` over `geom_boxplot` at small n.\n\n## UpSet for Multi-set Comparisons\n\nFor >3 DE gene sets (e.g., contrasts treated_drugA, treated_drugB, treated_drugC each vs control), Venn diagrams become unreadable. UpSet (Lex et al. 2014 *IEEE Trans Vis Comput Graph* 20:1983) scales:\n\n```r\nlibrary(UpSetR)\nupset(fromList(list(drugA = sig_drugA, drugB = sig_drugB, drugC = sig_drugC)))\n```\n\n## Per-Method Failure Modes\n\n### Volcano with unshrunken LFC -- inflated story\n\n**Trigger:** `ggplot(res_df, aes(x=log2FoldChange, ...))` without `lfcShrink()`; extreme dots at the corners are low-count genes.\n\n**Mechanism:** Unshrunken MLE LFCs are dominated by very-low-count genes whose log ratios are noisy. The visual top-left and top-right corners look impressive but are artifacts.\n\n**Symptom:** Top genes by abs(LFC) are obscure low-count genes; reviewer asks \"why are these the top hits?\"\n\n**Fix:** `res_apeglm \u003C- lfcShrink(dds, coef=..., type='apeglm')`; plot from `res_apeglm`. Label axis \"shrunken log2 fold change (apeglm)\".\n\n### ggrepel `max.overlaps` silently drops labels\n\n**Trigger:** `geom_text_repel(data = top30, aes(label = gene))`; only 10 labels render.\n\n**Mechanism:** Default `max.overlaps = 10`; warning printed but easily missed in a knitr\u002FQuarto render.\n\n**Symptom:** Reviewer asks \"where is gene X?\"; it was in `top30` but did not render.\n\n**Fix:** `geom_text_repel(..., max.overlaps = Inf)` or `options(ggrepel.max.overlaps = Inf)` at top of script.\n\n### PCA shows batch, not condition\n\n**Trigger:** `plotPCA(vsd, intgroup='batch')` cleanly separates batches; `intgroup='condition'` does not separate.\n\n**Mechanism:** Batch variance exceeds condition variance.\n\n**Symptom:** Treatment effect looks weak; DE p-values inflated if batch not in design.\n\n**Fix:** Include batch in design (`design = ~ batch + condition`). DO NOT use `removeBatchEffect` then re-do DE on corrected counts (Nygaard 2016 cardinal sin -- see `batch-correction`). For VISUALIZATION only, `removeBatchEffect` is OK.\n\n### Heatmap row-scaling hid a sample-level shift\n\n**Trigger:** QC heatmap with `scale='row'` looks consistent within group; downstream PCA shows clear sample outlier.\n\n**Mechanism:** z-score per gene removes per-sample additive shifts. A sample that's globally inflated 1.5x looks identical to peers after row scaling.\n\n**Symptom:** \"The heatmap looked fine but PCA shows a problem.\"\n\n**Fix:** For QC heatmaps, use `scale = 'none'` on `assay(vsd)` directly. For result heatmaps after QC is clean, `scale = 'row'` is the appropriate choice for pattern emphasis.\n\n### Top-N-by-rowVars dominated by single-outlier-sample genes\n\n**Trigger:** \"Top 500 variable genes\" PCA shows a striped pattern, one or two samples driving the spread.\n\n**Mechanism:** `rowVars` is squared-deviation; one outlier sample of one gene inflates that gene's \"variance\" massively.\n\n**Symptom:** Top variable gene list includes many genes where N-1 samples are flat and one sample is extreme.\n\n**Fix:** `matrixStats::rowMads()` for MAD-based selection; or `genefilter::rowQ()`.\n\n## Common errors\n\n| Error \u002F symptom | Cause | Fix |\n|-----------------|-------|-----|\n| `plotPCA` reports only 2 PCs | DESeq2 `plotPCA` is hard-coded to PC1\u002FPC2 | Use `prcomp(t(assay(vsd)))` and plot any pair |\n| PCA cloud collapses to one point | Forgot to log-transform; raw counts plotted | `vst(dds)` first |\n| All MA-plot points red | `alpha` set too high or sig-flag bug | Verify `alpha`; check `padj` vs `pvalue` in flag |\n| `pheatmap` complains \"infinite values\" | NA \u002F Inf in scaled matrix; gene with zero variance | Remove zero-variance rows before scaling |\n| Volcano axis labels obscured | Default ggplot theme too compact | `theme_bw(base_size = 14)` |\n| `plotCounts` says gene not found | Wrong ID type (symbol vs Ensembl) | Match `rownames(dds)` exactly |\n| `vst()` errors with very low gene count post-filter | Default `nsub=1000` exceeds available genes | Lower `nsub` (e.g., `vst(dds, nsub=500)`) |\n\n## References\n\n- Anders S, Huber W. 2010. Differential expression analysis for sequence count data. *Genome Biol* 11(10):R106. doi:10.1186\u002Fgb-2010-11-10-r106\n- Love MI, Huber W, Anders S. 2014. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. *Genome Biol* 15(12):550. doi:10.1186\u002Fs13059-014-0550-8\n- Zhu A, Ibrahim JG, Love MI. 2019. Heavy-tailed prior distributions for sequence count data: removing the noise and preserving large differences. *Bioinformatics* 35(12):2084-2092. doi:10.1093\u002Fbioinformatics\u002Fbty895\n- Robinson MD, McCarthy DJ, Smyth GK. 2010. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. *Bioinformatics* 26(1):139-140. doi:10.1093\u002Fbioinformatics\u002Fbtp616\n- Lex A, Gehlenborg N, Strobelt H, Vuillemot R, Pfister H. 2014. UpSet: Visualization of Intersecting Sets. *IEEE Trans Vis Comput Graph* 20(12):1983-1992. doi:10.1109\u002FTVCG.2014.2346248\n- Schurch NJ et al. 2016. How many biological replicates are needed in an RNA-seq experiment and which differential expression tool should you use? *RNA* 22(6):839-851. doi:10.1261\u002Frna.053959.115\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(1):29-39. doi:10.1093\u002Fbiostatistics\u002Fkxv027\n\n## Related Skills\n\n- deseq2-basics - Generates the `dds` \u002F `res` objects plotted here; `vst`\u002F`rlog` choice\n- edger-basics - Generates `y` \u002F `qlf` for plotMD, plotBCV, plotMDS\n- de-results - p-value histogram, padj=NA diagnosis informs what to plot\n- batch-correction - removeBatchEffect for visualization only (never as DE input)\n- expression-matrix\u002Fnormalization - VST vs rlog vs log-CPM mechanics\n- data-visualization\u002Fvolcano-and-ma-plots - Full custom volcano\u002FMA with apeglm + ggrepel\n- data-visualization\u002Fdimensionality-reduction-plots - PCA, UMAP, t-SNE customization\n- data-visualization\u002Fheatmaps-clustering - pheatmap and ComplexHeatmap recipes\n- data-visualization\u002Fupset-plots - UpSet plot customization\n","## Version Compatibility\n\nReference examples tested with: DESeq2 1.42+, edgeR 4.0+, limma 3.58+, ggplot2 3.5+, pheatmap 1.0+, RColorBrewer 1.1+, ggrepel 0.9+, EnhancedVolcano 1.20+, matrixStats 1.2+\n\nBefore using code patterns, verify installed versions match. If versions differ:\n- R: `packageVersion('\u003Cpkg>')` then `?function_name` to verify parameters\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\n# DE Visualization\n\n**\"Make the standard DE figure panel\"** -> Use built-in functions or thin wrappers to produce diagnostic plots (dispersion, p-value histogram, PCA, sample distance) and result plots (MA, volcano, heatmap of top DE genes, per-gene counts), interpreted as diagnostics of the underlying model.\n\n## Scope\n\nThis skill covers DE-specific built-in plots and immediate wrappers. For richer customization:\n- Custom volcano\u002FMA with apeglm-shrunken LFC and ggrepel labelling -> `data-visualization\u002Fvolcano-and-ma-plots`\n- PCA \u002F UMAP \u002F t-SNE customization -> `data-visualization\u002Fdimensionality-reduction-plots`\n- Heatmap customization and ComplexHeatmap recipes -> `data-visualization\u002Fheatmaps-clustering`\n\n## The Single Most Important Modern Insight -- A volcano with shrunken LFC compresses the cloud, but the p-values are unchanged\n\n`lfcShrink()` pulls noisy estimates toward zero. On the volcano, that pulls genes horizontally toward the center. But the y-axis (`-log10(pvalue)`) is the **unshrunken Wald p-value** -- shrinkage does NOT recompute p-values (Zhu, Ibrahim, Love 2019 *Bioinformatics* 35:2084). A naive reader sees fewer extreme dots and concludes \"fewer genes are significant\". Wrong: the same genes are significant; the effect sizes are smaller and more honest.\n\nAlways label the volcano x-axis \"shrunken log2 fold change (apeglm)\" and note the y-axis comes from the unshrunken Wald test. The whole point of the apeglm volcano is the honest effect-size axis; if a publication shows an unshrunken volcano, it is showing inflated effects from low-count noise.\n\nThe MA plot has its own version of this: shrinkage flattens the left side (low-mean, formerly extreme LFC) and barely touches the right (high-mean, well-estimated LFC). That asymmetry is the visual signature of working shrinkage.\n\n## Plot Taxonomy\n\n| Plot | Diagnostic OR result | Built-in function | What it tests |\n|------|---------------------|-------------------|---------------|\n| Dispersion plot | Diagnostic | `plotDispEsts(dds)` (DESeq2), `plotBCV(y)` (edgeR) | Mean-dispersion trend fit quality |\n| p-value histogram | Diagnostic | None; use ggplot2 | Null calibration, hidden batch, over-correction |\n| PCA on VST\u002Frlog | Diagnostic + result | `plotPCA(vsd, intgroup=...)` (DESeq2), `plotMDS()` (edgeR via limma) | Sample clustering, batch effects, outliers |\n| Sample distance heatmap | Diagnostic | `pheatmap` on `dist(t(assay(vsd)))` | Within-group consistency, sample swaps |\n| MA plot | Diagnostic + result | `plotMA(res)` (DESeq2), `plotMD(qlf)` (edgeR) | Normalization sanity, LFC vs mean |\n| Volcano | Result | ggplot2 wrapper; `EnhancedVolcano` | Top-effect, top-significance gene story |\n| Top-DE heatmap | Result | `pheatmap` on `assay(vsd)[sig_genes,]` | Per-gene pattern across conditions |\n| `plotCounts` per gene | Result | `plotCounts(dds, gene, intgroup)` | Per-gene biology |\n\n## Decision Tree by Scenario\n\n| Scenario | Recommended approach |\n|----------|---------------------|\n| PCA for unbiased QC | `vst(dds, blind = TRUE)`; ask \"do samples group as expected without design influence?\" |\n| PCA for results figure | `vst(dds, blind = FALSE)`; design is settled, accept its influence on dispersion |\n| n \u003C 30, library sizes vary >4x | `rlog(dds, blind = FALSE)` instead of vst |\n| n > 30 | `vst()`; rlog impractical |\n| Volcano | Plot shrunken LFC on x, unshrunken p-value on y; label both axes |\n| Sample distance heatmap | `vst(blind = TRUE)`; tells if a sample is the wrong group regardless of design |\n| Top-DE heatmap, want to see PATTERN | `scale = 'row'` (z-score per gene) |\n| Top-DE heatmap, want to see ABSOLUTE LEVEL | `scale = 'none'` on `assay(vsd)`; otherwise weak signal looks strong |\n| Top-variable-gene selection | `matrixStats::rowMads(assay(vsd))` instead of `rowVars` -- MAD is outlier-robust |\n| n = 3, top genes in volcano | Note Schurch 2016 finding: 20-40% of true positives missed; treat as exploratory |\n| Many groups, comparing DE sets | UpSet plot (Lex 2014); Venn drowns above 3 sets |\n\n## Dispersion Diagnostic (Run This First)\n\n**Goal:** Verify the dispersion-mean trend was fit acceptably before trusting any results.\n\n**Approach:** `plotDispEsts(dds)` (DESeq2) or `plotBCV(y)` (edgeR) shows gene-wise (black\u002Fblue), fitted trend (red), and final shrunken (blue) dispersions vs mean.\n\n```r\nplotDispEsts(dds)\n\nplotBCV(y)\n```\n\n| Pattern | Meaning | Action |\n|---------|---------|--------|\n| Cloud follows trend; final shrunken estimates pulled toward red curve | Healthy fit | Proceed |\n| Red trend nowhere near the gene-wise cloud | Parametric trend failed | `DESeq(dds, fitType = 'local')` or `fitType = 'mean'` |\n| Many gene-wise dispersions FAR ABOVE the trend | Outlier or unmodeled batch genes | Inspect rather than trust QL F-test alone |\n| Final estimates much lower than gene-wise everywhere | Excessive shrinkage; sample too small or trend too flat | Check `useEM`, robust hyperparameter setting |\n| BCV decreases monotonically with mean | Correct in edgeR | Default trend |\n\nA plot inspected before trusting results is worth a hundred lines of statistical safeguards.\n\n## P-value Histogram (Run This Second)\n\n**Goal:** Detect model misspecification or hidden batch before reporting any gene list.\n\n**Approach:** Histogram of raw p-values; under a correctly specified null, uniform with a spike near zero.\n\n```r\nlibrary(ggplot2)\nggplot(res_df, aes(x = pvalue)) +\n    geom_histogram(bins = 50, fill = 'steelblue', color = 'white') +\n    labs(x = 'P-value', y = 'Frequency', title = 'P-value distribution') +\n    theme_bw()\n```\n\n| Shape | Meaning | Action |\n|-------|---------|--------|\n| Uniform + spike at 0 | Correctly specified | Proceed |\n| U-shape (spikes at 0 AND 1) | Anti-conservative; hidden batch or unmodeled covariate | Add the missing covariate; re-fit |\n| Depleted near 0, spike near 1 | Conservative; over-modeled or wrong dispersion | Simplify model; check dispersion plot |\n| Spike only at p = 1 | Discrete artifact from very-low-count genes | Pre-filter more aggressively |\n\n## MA Plot (LFC vs Mean)\n\n**Goal:** Inspect the relationship between LFC and mean expression for normalization correctness and shrinkage effect.\n\n**Approach:** `plotMA` (DESeq2) or `plotMD` (edgeR). Always pick `ylim` deliberately; default can flatten the signal.\n\n```r\nplotMA(res, ylim = c(-5, 5), main = 'MA plot (unshrunken)')\n\nres_apeglm \u003C- lfcShrink(dds, coef = 'condition_treated_vs_control', type = 'apeglm')\nplotMA(res_apeglm, ylim = c(-5, 5), main = 'MA plot (apeglm-shrunken)')\n\nplotMD(qlf, main = 'edgeR MD plot')\nabline(h = c(-1, 1), col = 'blue', lty = 2)\n```\n\n| Pattern | Meaning |\n|---------|---------|\n| Symmetric cloud centered at LFC = 0 | Correct normalization |\n| Cloud median clearly above or below 0 | Normalization failed (TMM\u002FRLE assumption violated) -- see normalization skill |\n| Funnel widening at low mean | Expected (low counts noisier) |\n| Dramatic up\u002Fdown asymmetry | Possibly real (large biological perturbation), possibly normalization failure -- cross-check |\n| Discrete horizontal bands at low mean | Low-count artifacts; pre-filter more aggressively |\n\nThe apeglm-shrunken MA visually flattens the left side; the post-shrinkage cloud should be tighter at low means.\n\n## Volcano with Shrunken LFC\n\n**Goal:** Show effect size vs significance with honest fold changes.\n\n**Approach:** Use a built-in renderer (EnhancedVolcano for quick publication-quality output) on shrunken LFCs. Always plot shrunken LFC; always set `max.overlaps = Inf` when labeling >10 genes -- the ggrepel default (10) silently drops labels. EnhancedVolcano accepts `max.overlaps` directly in 1.12+; version 1.10-1.11 has the older `maxoverlapsConnectors` argument (default 15); for either, falling back to `options(ggrepel.max.overlaps = Inf)` at the top of the script also works. For full ggplot2 customization (color schemes, faceting, label-set engineering), see `data-visualization\u002Fvolcano-and-ma-plots`.\n\n```r\nlibrary(EnhancedVolcano)\n\nres_apeglm \u003C- lfcShrink(dds, coef = 'condition_treated_vs_control', type = 'apeglm')\n\nEnhancedVolcano(res_apeglm,\n    lab = rownames(res_apeglm),\n    x = 'log2FoldChange', y = 'pvalue',\n    pCutoff = 0.05, FCcutoff = 1,\n    title = 'Treatment vs Control',\n    subtitle = 'Shrunken LFC (apeglm); unshrunken Wald p',\n    max.overlaps = Inf)\n```\n\n## PCA on VST\u002Frlog (Never on Raw Counts)\n\n**Goal:** Show sample clustering by condition; detect batch effects, swaps, outliers.\n\n**Approach:** Variance-stabilize first (VST or rlog), THEN PCA. Raw counts make PC1 = library size; log(counts+1) makes PC1 = mean expression. Neither carries biological signal until variance is stabilized.\n\n```r\nvsd \u003C- vst(dds, blind = FALSE)\nplotPCA(vsd, intgroup = c('condition', 'batch'))\n\npca_df \u003C- plotPCA(vsd, intgroup = c('condition', 'batch'), returnData = TRUE)\npercentVar \u003C- round(100 * attr(pca_df, 'percentVar'))\n\nlibrary(ggplot2)\nggplot(pca_df, aes(PC1, PC2, color = condition, shape = batch)) +\n    geom_point(size = 4) +\n    xlab(paste0('PC1: ', percentVar[1], '% variance')) +\n    ylab(paste0('PC2: ', percentVar[2], '% variance')) +\n    theme_bw()\n\nlibrary(limma)\nplotMDS(cpm(y, log = TRUE), col = as.numeric(group), pch = 16)\n```\n\n`blind=TRUE` (default for `vst()`) re-estimates dispersions ignoring the design -- appropriate for unbiased QC (\"are samples consistent independent of design?\"). `blind=FALSE` uses the fitted dispersions -- appropriate for downstream visualization where the design is settled. Modern DESeq2 vignette recommends `blind=FALSE` for any plot after the model is fit.\n\n| PCA pattern | Interpretation | Action |\n|-------------|----------------|--------|\n| Clear separation by condition on PC1 or PC2 | Strong biological signal | Proceed |\n| Separation by batch, not condition | Batch effect dominates | Include batch in design; DO NOT subtract before DE (see batch-correction Nygaard 2016) |\n| One sample far from its group | Outlier or swap | Check library QC; sex check; somalier |\n| Condition signal on PC3+, not PC1-PC2 | Subtle effect | May still find DE; review dispersion plot |\n| Two distinct sample clusters not explained by metadata | Hidden covariate | Investigate processing date, lane, machine |\n\n## Sample Distance Heatmap (for QC)\n\n```r\nlibrary(pheatmap)\nvsd \u003C- vst(dds, blind = TRUE)\nsd \u003C- dist(t(assay(vsd)))\nmat \u003C- as.matrix(sd)\nann \u003C- data.frame(condition = colData(dds)$condition,\n                  row.names = colnames(dds))\npheatmap(mat, annotation_col = ann, annotation_row = ann,\n         clustering_distance_rows = sd, clustering_distance_cols = sd,\n         color = colorRampPalette(c('white', 'steelblue'))(100),\n         main = 'Sample distance (vst blind)')\n```\n\nThe diagonal should be dark; within-group samples should cluster. A within-group sample distant from its peers is a candidate for sample swap.\n\n## Top-DE Heatmap and the Row-Scaling Trap\n\n**Goal:** Show expression patterns of significant genes across samples for results figure.\n\n**Approach:** Use `vst(blind=FALSE)`, select top genes (by adjusted p-value or MAD-robust variance), choose scaling deliberately.\n\n```r\nlibrary(pheatmap)\n\nsig \u003C- rownames(subset(res, padj \u003C 0.01))[1:50]\nvsd \u003C- vst(dds, blind = FALSE)\nmat \u003C- assay(vsd)[sig, ]\n\nmat_scaled \u003C- t(scale(t(mat)))\n\nann_col \u003C- data.frame(condition = colData(dds)$condition,\n                      batch     = colData(dds)$batch,\n                      row.names = colnames(mat))\n\npheatmap(mat_scaled, annotation_col = ann_col,\n         show_rownames = FALSE,\n         clustering_distance_rows = 'correlation',\n         clustering_distance_cols = 'correlation',\n         color = colorRampPalette(c('blue', 'white', 'red'))(100),\n         main = 'Top 50 DE genes (z-scored per gene)')\n```\n\n`scale='row'` (z-score per gene) is the conventional choice for \"show me patterns\". It DESTROYS absolute expression level information -- a gene at 5-7 with mean 6 looks identical to a gene at 10-1000. For pattern detection: correct. For QC heatmaps showing batch shifts: WRONG -- use `scale='none'` on `assay(vsd)`.\n\nTop-variable-gene selection robustness:\n\n```r\nlibrary(matrixStats)\nvars_mad \u003C- rowMads(assay(vsd))\ntop500 \u003C- order(vars_mad, decreasing = TRUE)[1:500]\n```\n\n`rowMads` (median absolute deviation) is outlier-robust; `rowVars` is dominated by single-outlier-sample genes. For exploratory PCA of \"top variable genes\", MAD selection avoids artifacts.\n\n## Per-gene Plot\n\n```r\nplotCounts(dds, gene = 'GENE_NAME', intgroup = 'condition')\n\nd \u003C- plotCounts(dds, gene = 'GENE_NAME', intgroup = c('condition','batch'),\n                returnData = TRUE)\nlibrary(ggplot2)\nggplot(d, aes(x = condition, y = count, color = batch)) +\n    geom_jitter(width = 0.1, size = 3) +\n    scale_y_log10() +\n    ggtitle('GENE_NAME') +\n    theme_bw()\n```\n\nWith n=3, the boxplot is misleading (3 points per box). Prefer `geom_jitter` over `geom_boxplot` at small n.\n\n## UpSet for Multi-set Comparisons\n\nFor >3 DE gene sets (e.g., contrasts treated_drugA, treated_drugB, treated_drugC each vs control), Venn diagrams become unreadable. UpSet (Lex et al. 2014 *IEEE Trans Vis Comput Graph* 20:1983) scales:\n\n```r\nlibrary(UpSetR)\nupset(fromList(list(drugA = sig_drugA, drugB = sig_drugB, drugC = sig_drugC)))\n```\n\n## Per-Method Failure Modes\n\n### Volcano with unshrunken LFC -- inflated story\n\n**Trigger:** `ggplot(res_df, aes(x=log2FoldChange, ...))` without `lfcShrink()`; extreme dots at the corners are low-count genes.\n\n**Mechanism:** Unshrunken MLE LFCs are dominated by very-low-count genes whose log ratios are noisy. The visual top-left and top-right corners look impressive but are artifacts.\n\n**Symptom:** Top genes by abs(LFC) are obscure low-count genes; reviewer asks \"why are these the top hits?\"\n\n**Fix:** `res_apeglm \u003C- lfcShrink(dds, coef=..., type='apeglm')`; plot from `res_apeglm`. Label axis \"shrunken log2 fold change (apeglm)\".\n\n### ggrepel `max.overlaps` silently drops labels\n\n**Trigger:** `geom_text_repel(data = top30, aes(label = gene))`; only 10 labels render.\n\n**Mechanism:** Default `max.overlaps = 10`; warning printed but easily missed in a knitr\u002FQuarto render.\n\n**Symptom:** Reviewer asks \"where is gene X?\"; it was in `top30` but did not render.\n\n**Fix:** `geom_text_repel(..., max.overlaps = Inf)` or `options(ggrepel.max.overlaps = Inf)` at top of script.\n\n### PCA shows batch, not condition\n\n**Trigger:** `plotPCA(vsd, intgroup='batch')` cleanly separates batches; `intgroup='condition'` does not separate.\n\n**Mechanism:** Batch variance exceeds condition variance.\n\n**Symptom:** Treatment effect looks weak; DE p-values inflated if batch not in design.\n\n**Fix:** Include batch in design (`design = ~ batch + condition`). DO NOT use `removeBatchEffect` then re-do DE on corrected counts (Nygaard 2016 cardinal sin -- see `batch-correction`). For VISUALIZATION only, `removeBatchEffect` is OK.\n\n### Heatmap row-scaling hid a sample-level shift\n\n**Trigger:** QC heatmap with `scale='row'` looks consistent within group; downstream PCA shows clear sample outlier.\n\n**Mechanism:** z-score per gene removes per-sample additive shifts. A sample that's globally inflated 1.5x looks identical to peers after row scaling.\n\n**Symptom:** \"The heatmap looked fine but PCA shows a problem.\"\n\n**Fix:** For QC heatmaps, use `scale = 'none'` on `assay(vsd)` directly. For result heatmaps after QC is clean, `scale = 'row'` is the appropriate choice for pattern emphasis.\n\n### Top-N-by-rowVars dominated by single-outlier-sample genes\n\n**Trigger:** \"Top 500 variable genes\" PCA shows a striped pattern, one or two samples driving the spread.\n\n**Mechanism:** `rowVars` is squared-deviation; one outlier sample of one gene inflates that gene's \"variance\" massively.\n\n**Symptom:** Top variable gene list includes many genes where N-1 samples are flat and one sample is extreme.\n\n**Fix:** `matrixStats::rowMads()` for MAD-based selection; or `genefilter::rowQ()`.\n\n## Common errors\n\n| Error \u002F symptom | Cause | Fix |\n|-----------------|-------|-----|\n| `plotPCA` reports only 2 PCs | DESeq2 `plotPCA` is hard-coded to PC1\u002FPC2 | Use `prcomp(t(assay(vsd)))` and plot any pair |\n| PCA cloud collapses to one point | Forgot to log-transform; raw counts plotted | `vst(dds)` first |\n| All MA-plot points red | `alpha` set too high or sig-flag bug | Verify `alpha`; check `padj` vs `pvalue` in flag |\n| `pheatmap` complains \"infinite values\" | NA \u002F Inf in scaled matrix; gene with zero variance | Remove zero-variance rows before scaling |\n| Volcano axis labels obscured | Default ggplot theme too compact | `theme_bw(base_size = 14)` |\n| `plotCounts` says gene not found | Wrong ID type (symbol vs Ensembl) | Match `rownames(dds)` exactly |\n| `vst()` errors with very low gene count post-filter | Default `nsub=1000` exceeds available genes | Lower `nsub` (e.g., `vst(dds, nsub=500)`) |\n\n## References\n\n- Anders S, Huber W. 2010. Differential expression analysis for sequence count data. *Genome Biol* 11(10):R106. doi:10.1186\u002Fgb-2010-11-10-r106\n- Love MI, Huber W, Anders S. 2014. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. *Genome Biol* 15(12):550. doi:10.1186\u002Fs13059-014-0550-8\n- Zhu A, Ibrahim JG, Love MI. 2019. Heavy-tailed prior distributions for sequence count data: removing the noise and preserving large differences. *Bioinformatics* 35(12):2084-2092. doi:10.1093\u002Fbioinformatics\u002Fbty895\n- Robinson MD, McCarthy DJ, Smyth GK. 2010. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. *Bioinformatics* 26(1):139-140. doi:10.1093\u002Fbioinformatics\u002Fbtp616\n- Lex A, Gehlenborg N, Strobelt H, Vuillemot R, Pfister H. 2014. UpSet: Visualization of Intersecting Sets. *IEEE Trans Vis Comput Graph* 20(12):1983-1992. doi:10.1109\u002FTVCG.2014.2346248\n- Schurch NJ et al. 2016. How many biological replicates are needed in an RNA-seq experiment and which differential expression tool should you use? *RNA* 22(6):839-851. doi:10.1261\u002Frna.053959.115\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(1):29-39. doi:10.1093\u002Fbiostatistics\u002Fkxv027\n\n## Related Skills\n\n- deseq2-basics - Generates the `dds` \u002F `res` objects plotted here; `vst`\u002F`rlog` choice\n- edger-basics - Generates `y` \u002F `qlf` for plotMD, plotBCV, plotMDS\n- de-results - p-value histogram, padj=NA diagnosis informs what to plot\n- batch-correction - removeBatchEffect for visualization only (never as DE input)\n- expression-matrix\u002Fnormalization - VST vs rlog vs log-CPM mechanics\n- data-visualization\u002Fvolcano-and-ma-plots - Full custom volcano\u002FMA with apeglm + ggrepel\n- data-visualization\u002Fdimensionality-reduction-plots - PCA, UMAP, t-SNE customization\n- data-visualization\u002Fheatmaps-clustering - pheatmap and ComplexHeatmap recipes\n- data-visualization\u002Fupset-plots - UpSet plot customization\n",{"name":5,"source":24,"license":15,"tool_type":25,"description":7,"imported_at":26,"source_path":27,"primary_tool":28,"source_repository":29},"https:\u002F\u002Fgithub.com\u002FGPTomics\u002FbioSkills","r","2026-09-20","differential-expression\u002Fde-visualization\u002FSKILL.md","DESeq2","GPTomics\u002FbioSkills","main","2d1750717747097b602acf30bd09a7cbd4b4c6ea","awesome_skills\u002Fbiology\u002Fbio-differential-expression-de-visualization","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":37,"prompt":47,"zipAvailable":48},[38,41,44],{"tool":39,"command":40},"npx","npx skills add Mr-Nobody-Anonymous\u002FAll-skills --skill bio-differential-expression-de-visualization",{"tool":42,"command":43},"bunx","bunx skills add Mr-Nobody-Anonymous\u002FAll-skills --skill bio-differential-expression-de-visualization",{"tool":45,"command":46},"pnpm","pnpm dlx skills add Mr-Nobody-Anonymous\u002FAll-skills --skill bio-differential-expression-de-visualization","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fbiology\u002Fbio-differential-expression-de-visualization\u002FSKILL.md",true,[50,56,60,64],{"path":51,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":54,"sourceHtmlUrl":55},"examples\u002Fheatmap.R","OTHER","其他",1335,"https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fbiology\u002Fbio-differential-expression-de-visualization\u002Fexamples\u002Fheatmap.R",{"path":57,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":58,"sourceHtmlUrl":59},"examples\u002Fpca_plot.R",1481,"https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fbiology\u002Fbio-differential-expression-de-visualization\u002Fexamples\u002Fpca_plot.R",{"path":61,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":62,"sourceHtmlUrl":63},"examples\u002Fvolcano_plot.R",1809,"https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fbiology\u002Fbio-differential-expression-de-visualization\u002Fexamples\u002Fvolcano_plot.R",{"path":65,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":68,"sourceHtmlUrl":69},"usage-guide.md","REFERENCE","参考文档",3537,"https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fbiology\u002Fbio-differential-expression-de-visualization\u002Fusage-guide.md",[71,79,86,93,100,107],{"skillNo":72,"name":73,"nameZh":6,"description":74,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":75,"categories":76,"keywords":77,"publishedAt":78},"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":80,"name":81,"nameZh":6,"description":82,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":83,"categories":84,"keywords":85,"publishedAt":78},"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":87,"name":88,"nameZh":6,"description":89,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":90,"categories":91,"keywords":92,"publishedAt":78},"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":94,"name":95,"nameZh":6,"description":96,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":97,"categories":98,"keywords":99,"publishedAt":78},"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":101,"name":102,"nameZh":6,"description":103,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":104,"categories":105,"keywords":106,"publishedAt":78},"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":108,"name":109,"nameZh":6,"description":110,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":111,"categories":112,"keywords":113,"publishedAt":78},"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",[],[]]