[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7ED38DYP7MB9TBVXGQDTMYN":3,"skill-same:01M7ED38DYP7MB9TBVXGQDTMYN":62},{"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},"01M7ED38DYP7MB9TBVXGQDTMYN","bio-epitranscriptomics-m6a-differential",null,"Identifies differential m6A methylation between conditions from MeRIP-seq paired IP\u002Finput data using exomePeak2 (GC-bias-aware differential via its bam_ip\u002Fbam_input control + bam_treated_ip\u002Fbam_treated_input treatment arms), QNB beta-binomial, MeTDiff HMM, and RADAR, plus the paired-symmetric edgeR\u002FDESeq2-on-peak-counts route when batch\u002Flot covariates need fixed-effect handling that exomePeak2's API does not accept. Covers paired vs unpaired vs interaction designs, batch confounding and per-lot meta-analysis, the stoichiometry-vs-expression-vs-IP-efficiency confound, and effect-size filtering against under-powered N=2 designs. Use when comparing m6A across two or more conditions, choosing between exomePeak2\u002FQNB\u002FRADAR\u002FMeTDiff for a design, handling batch confounding when exomePeak2's API is too rigid, distinguishing real hyper\u002Fhypo-methylation from expression shifts, applying effect-size thresholds, or planning orthogonal stoichiometry validation (GLORI\u002FSAC-seq\u002Fm6Anet mod_ratio).","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-epitranscriptomics-m6a-differential\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: bio-epitranscriptomics-m6a-differential\ndescription: Identifies differential m6A methylation between conditions from MeRIP-seq paired IP\u002Finput data using exomePeak2 (GC-bias-aware differential via its bam_ip\u002Fbam_input control + bam_treated_ip\u002Fbam_treated_input treatment arms), QNB beta-binomial, MeTDiff HMM, and RADAR, plus the paired-symmetric edgeR\u002FDESeq2-on-peak-counts route when batch\u002Flot covariates need fixed-effect handling that exomePeak2's API does not accept. Covers paired vs unpaired vs interaction designs, batch confounding and per-lot meta-analysis, the stoichiometry-vs-expression-vs-IP-efficiency confound, and effect-size filtering against under-powered N=2 designs. Use when comparing m6A across two or more conditions, choosing between exomePeak2\u002FQNB\u002FRADAR\u002FMeTDiff for a design, handling batch confounding when exomePeak2's API is too rigid, distinguishing real hyper\u002Fhypo-methylation from expression shifts, applying effect-size thresholds, or planning orthogonal stoichiometry validation (GLORI\u002FSAC-seq\u002Fm6Anet mod_ratio).\ntool_type: r\nprimary_tool: exomePeak2\nsource: \"https:\u002F\u002Fgithub.com\u002FGPTomics\u002FbioSkills\"\nsource_repository: \"GPTomics\u002FbioSkills\"\nsource_path: \"epitranscriptomics\u002Fm6a-differential\u002FSKILL.md\"\nlicense: \"MIT\"\nimported_at: \"2026-09-20\"\n---\n\n## Version Compatibility\n\nReference examples tested with: exomePeak2 1.14+ (Bioconductor 3.18+), QNB 1.1.11 (GitHub `lzcyzm\u002FQNB`), MeTDiff (GitHub, bundled with MeTPeak), RADAR 0.2.4+ (GitHub `scottzijiezhang\u002FRADAR`), DESeq2 1.42+, edgeR 4.0+, GenomicFeatures 1.54+, ggplot2 3.5+, GenomicAlignments 1.38+, Rsubread 2.16+.\n\nBefore using code patterns, verify installed versions match. If versions differ:\n- R: `packageVersion('exomePeak2')` then `?exomePeak2` to verify parameters\n- R: `packageVersion('QNB')` then `?qnbtest` to confirm argument signature\n\nIf R throws `unused argument` or `argument is missing`, the API moved between Bioconductor minor releases; consult `?exomePeak2` directly. QNB function signature has been stable since 2017 but the GitHub source has occasional changes; pin the commit SHA.\n\nexomePeak2's differential interface is NOT a `mode=` argument — populate the `bam_treated_ip=` and `bam_treated_input=` arguments alongside the standard `bam_ip=` \u002F `bam_input=` (control arm) to trigger paired differential calling. `peak_calling_mode` is a separate argument controlling locus scope (`exon` | `full_transcript` | `whole_genome`). Verify against `?exomePeak2` and Bioconductor 3.20+ release notes. QNB is on GitHub only; install via `devtools::install_github('lzcyzm\u002FQNB')`. RADAR is on GitHub only; install via `devtools::install_github('scottzijiezhang\u002FRADAR')`; its differential workflow is `countReads -> normalizeLibrary -> adjustExprLevel -> filterBins -> diffIP -> reportResult`.\n\n# Differential m6A Analysis\n\n**\"Compare m6A methylation between my conditions\"** -> Quantify how much each m6A peak's IP\u002Finput enrichment shifts between conditions, after normalising for transcript-abundance changes (which all show up in input) and for IP-efficiency drift (which the design matrix and within-run replicates control for). Then apply effect-size filtering to distinguish real biology from the technical noise floor that all MeRIP differential methods inherit (McIntyre 2020 *Sci Rep* 10:6590: between-study m6A peak overlap is ~45% median; differential calls within that noise envelope routinely fail to replicate). Critically: a higher MeRIP signal in condition A vs B can mean (1) more transcripts of the peak-bearing gene, (2) more methylation per transcript, OR (3) higher IP efficiency in batch A — distinguishing requires careful normalisation or an orthogonal absolute-stoichiometry method.\n\n- R: `exomePeak2::exomePeak2(bam_ip, bam_input, bam_treated_ip, bam_treated_input, ...)` -- integrated peak + differential GLM (modern default)\n- R: `QNB::qnbtest(control_ip, treated_ip, control_input, treated_input)` -- beta-binomial for small N (Liu 2017 *BMC Bioinformatics* 18:387)\n- R: `RADAR::diffIP()` then `reportResult()` -- Poisson-NB on peak windows with TMM normalisation (Zhang 2019 *Genome Biol* 20:294)\n- R: `MeTDiff::metdiff()` -- HMM-based differential paired with MeTPeak\n- R: edgeR \u002F DESeq2 on featureCounts-on-peaks matrix -- defensible for paired symmetric designs with strong input normalisation; also the route for arbitrary batch \u002F lot covariates exomePeak2's API does not accept\n\n## The Single Most Important Modern Insight -- IP fold-change between conditions conflates stoichiometry change with expression change and IP efficiency drift\n\nA higher m6A peak signal in condition A vs B can mean ANY of: (1) more transcripts of the peak-bearing gene (expression up; the input increases proportionally, so the ratio should not change — but residual normalisation noise leaks through), (2) more methylation per transcript (stoichiometry up; the real biology of interest), (3) higher IP efficiency in batch A (technical; antibody lot, IP day, RNA prep). Differential MeRIP WITHOUT per-window input-normalisation OR an orthogonal stoichiometry-aware method (GLORI Liu 2023 *Nat Biotechnol* 41:355; SAC-seq Hu 2022 *Nat Biotechnol* 40:1210; MAZTER-seq Garcia-Campos 2019 *Cell* 178:731; m6Anet per-read modification rate) cannot separate the three. exomePeak2 differential mode, QNB, RADAR, and MeTDiff all implement per-window IP\u002Finput ratio modelling, but each has different default normalisations and the choice matters. Equally critical: McIntyre 2020 *Sci Rep* 10:6590 showed that \"differential\" m6A peaks from MeRIP-seq routinely do not replicate between independent studies in nominally identical conditions. The empirical noise floor is high; effect-size filtering (|log2FC| >= 0.5 minimum, often >= 1) AND replicate-direction concordance (the change is consistent in direction across replicates) AND minimum N=3 per condition are needed for differential calls to survive replication. For any absolute stoichiometry claim (\"this peak is 80% methylated in tumour vs 20% in normal\"), require an orthogonal stoichiometry method, not MeRIP alone.\n\n## Algorithmic Taxonomy\n\n| Tool \u002F mode | Mechanism | Inputs | Output | Strength | Fails when |\n|-------------|-----------|--------|--------|----------|------------|\n| exomePeak2 differential (Liu 2022) | Transcript-windowed Poisson GLM with GC-bias correction; integrated peak + differential | control IP\u002Finput BAM vectors + treated IP\u002Finput BAM vectors + TxDb + BSgenome | Differential peaks with log2FC + FDR per peak | Most use cases; integrates with peak calling; modern default | Small-N (\u003C=2) overdispersion poorly estimated; top-level API does NOT accept arbitrary covariates (use DESeq2\u002FedgeR for batch-aware designs) |\n| QNB (Liu 2017 *BMC Bioinformatics* 18:387) | Quad-negative-binomial joint model of IP, input, condition | per-peak count matrices (4 matrices: ip1, ip2, input1, input2) | Differential peaks with p-value + log2 RR | Designed for small N (2-3 per group); handles overdispersion explicitly | Requires pre-computed count matrices; not integrated with peak calling |\n| MeTDiff (bundled with MeTPeak) | HMM + Beta-binomial differential paired with MeTPeak | paired IP\u002Finput BAM + GTF + condition factor | Differential peaks per window | Pairs naturally with MeTPeak output; HMM smoothing helps low-coverage | GitHub-only; less benchmarked than exomePeak2 \u002F QNB |\n| RADAR (Zhang 2019 *Genome Biol* 20:294) | Poisson-NB with TMM normalisation; reproducibility-aware | paired IP\u002Finput BAM + condition factor | Differential peaks with logFC + FDR | Explicit replicate-variance modeling; reproducibility-aware framework | GitHub-only; slower than exomePeak2 |\n| DRME (Liu 2016 *Anal Biochem* 499:15) | Count-based small-sample alternative | per-peak count matrices | Differential peaks | Sister to QNB from same group; small-N alternative | Less benchmarked than QNB \u002F exomePeak2 |\n| edgeR \u002F DESeq2 on peak counts | Generic RNA-seq differential framework applied to featureCounts-on-peaks | peak count matrix + sample sheet | Differential peaks with log2FC + FDR | Familiar; flexible designs; well-tested in RNA-seq | Treats peak counts as RNA counts; loses IP\u002Finput pairing structure; defensible only for paired symmetric designs with strong input normalisation |\n| Ratio-of-ratios (heuristic) | Compute per-peak log2 (IP_A \u002F Input_A) - log2 (IP_B \u002F Input_B) per sample, then t-test | per-peak count matrix | per-peak t-test | Transparent; no model assumptions | No multiple-testing correction; ignores overdispersion; not recommended for primary analysis |\n\n## Decision Tree by Scenario\n\n| Scenario | Recommended | Why wrong choices fail |\n|----------|-------------|------------------------|\n| Standard 3-vs-3 paired-design MeRIP differential | exomePeak2 with `bam_ip` (ctrl) + `bam_treated_ip` (treat) and matching inputs | QNB usable but designed for smaller N; edgeR\u002FDESeq2 loses IP\u002Finput pairing |\n| Very small N (2 vs 2) | QNB (designed for small-sample overdispersion); supplement with exomePeak2 if possible | edgeR \u002F DESeq2 dispersion estimation collapses; exomePeak2 GLM also struggles at N=2 |\n| Paired design (patient as blocking factor) | QNB per-pair then aggregate; OR featureCounts-on-peaks -> DESeq2 with `~patient + condition` design; exomePeak2 cannot encode patient blocking via its top-level API | Unpaired analysis inflates within-group variance |\n| Interaction design (genotype × treatment) | featureCounts-on-peaks -> DESeq2 \u002F edgeR with interaction term; exomePeak2 top-level API is two-group only | QNB pairwise only; build interaction model from pairwise contrasts manually |\n| Batch confounding (antibody lot, sequencing run, IP day) | Include batch as fixed effect in DESeq2 \u002F edgeR model on featureCounts-on-peaks matrix; OR run exomePeak2 per-batch and meta-analyse | Pooling cross-batch counts without batch term attributes lot-effect to condition |\n| Time-course differential | featureCounts-on-peaks -> DESeq2 \u002F limma with time as numeric covariate; OR pairwise time-point exomePeak2 contrasts | Naive group-vs-group ignores time structure |\n| Stoichiometry claims (not just enrichment) | NOT MeRIP differential -- orthogonal GLORI \u002F SAC-seq \u002F m6Anet per-read | MeRIP IP fold-change is relative; cannot give per-molecule stoichiometry |\n| Cross-batch differential (different antibody lots) | featureCounts-on-peaks -> DESeq2 \u002F edgeR with `lot` in design; OR run exomePeak2 per-lot then meta-analyse; ideally avoid confounding lot with condition at the experimental-design stage | Lot effect inflates false positives; exomePeak2 top-level API cannot encode lot |\n| Validation of differential calls | Run >=2 differential methods; require concordant direction across replicates AND |log2FC| > 0.5 minimum (>= 1 stringent); orthogonal validation at top hits | Single-method differential calls within technical noise floor (per McIntyre 2020) |\n| Visualising differential peaks | Volcano plot with |log2FC| + FDR thresholds; MA plot to inspect normalisation; per-peak boxplot for top hits | Single number summaries hide stoichiometry vs expression confound |\n| Wanting to test a single gene \u002F locus | Targeted: per-peak boxplot across replicates with condition factor; manual t-test or Wilcoxon at high-coverage peak | Whole-transcriptome differential testing wastes multiple-testing budget for single-locus questions |\n\nMethodology evolves; before any high-stakes differential analysis, web-search \"exomePeak2 differential mode Bioconductor 3.20\" and \"MeRIP differential benchmark McIntyre\" for current consensus parameters.\n\n## exomePeak2 Differential Workflow\n\n**Goal:** Identify m6A peaks that differ in methylation level between conditions, controlling for transcript-abundance differences (via input normalisation) and GC bias (via internal correction), with an integrated peak-calling + differential pipeline.\n\n**Approach:** Build TxDb from the matched GTF; pass control IP\u002Finput BAM vectors via `bam_ip` and `bam_input` AND treatment IP\u002Finput BAM vectors via `bam_treated_ip` and `bam_treated_input` — populating the treated arms triggers differential mode (there is no separate `mode=` argument). Output is per-peak log2FC + FDR.\n\n```r\nlibrary(exomePeak2)\nlibrary(GenomicFeatures)\nlibrary(BSgenome.Hsapiens.UCSC.hg38)\n\ntxdb \u003C- makeTxDbFromGFF('refs\u002Fannotation.gtf', format='gtf')\n\nctrl_ip      \u003C- c('aligned\u002Fctrl_IP1.bam', 'aligned\u002Fctrl_IP2.bam', 'aligned\u002Fctrl_IP3.bam')\nctrl_input   \u003C- c('aligned\u002Fctrl_Input1.bam', 'aligned\u002Fctrl_Input2.bam', 'aligned\u002Fctrl_Input3.bam')\ntreat_ip     \u003C- c('aligned\u002Ftreat_IP1.bam', 'aligned\u002Ftreat_IP2.bam', 'aligned\u002Ftreat_IP3.bam')\ntreat_input  \u003C- c('aligned\u002Ftreat_Input1.bam', 'aligned\u002Ftreat_Input2.bam', 'aligned\u002Ftreat_Input3.bam')\n\nresult \u003C- exomePeak2(\n    bam_ip             = ctrl_ip,\n    bam_input          = ctrl_input,\n    bam_treated_ip     = treat_ip,\n    bam_treated_input  = treat_input,\n    txdb               = txdb,\n    genome             = BSgenome.Hsapiens.UCSC.hg38,\n    paired_end         = TRUE,\n    library_type       = 'unstranded',\n    peak_calling_mode  = 'exon',\n    save_dir           = 'exomepeak2_diff_output',\n    experiment_name    = 'ctrl_vs_treat'\n)\n\ndiff_table \u003C- as.data.frame(result)\nnrow(diff_table)\nhead(diff_table[, c('seqnames', 'start', 'end', 'log2FC', 'pvalue', 'padj')])\n```\n\n`peak_calling_mode` accepts `'exon'` (transcript-aware, default), `'full_transcript'`, or `'whole_genome'`; the meaning is locus scope, NOT differential-vs-non-differential. For arbitrary covariate adjustment (batch, antibody lot, patient blocking), the exomePeak2 top-level API is insufficient — move counts into DESeq2 \u002F edgeR via the featureCounts-on-peaks route below.\n\n## QNB Beta-Binomial for Small-N Designs\n\n**Goal:** Test differential m6A at pre-called peaks using a quad-negative-binomial model that handles small-N overdispersion better than generic GLM frameworks.\n\n**Approach:** Count reads in each IP and Input BAM at each peak using featureCounts or summarizeOverlaps; pass the four count matrices (ip1, ip2, input1, input2 — ip\u002Finput per group) to `qnbtest()`.\n\n```r\nlibrary(QNB)\nlibrary(Rsubread)\nlibrary(rtracklayer)\n\npeaks \u003C- import('exomepeak2_output\u002Fm6a_run1\u002Fpeaks.bed')\npeak_saf \u003C- data.frame(\n    GeneID = paste0('peak_', seq_along(peaks)),\n    Chr    = as.character(seqnames(peaks)),\n    Start  = start(peaks),\n    End    = end(peaks),\n    Strand = as.character(strand(peaks))\n)\n\ncount_matrix \u003C- function(bam_paths, peak_saf) {\n    fc \u003C- featureCounts(\n        files       = bam_paths,\n        annot.ext   = peak_saf,\n        isPairedEnd = TRUE,\n        nthreads    = 8,\n        allowMultiOverlap = TRUE\n    )\n    fc$counts\n}\n\nip_ctrl   \u003C- count_matrix(c('aligned\u002Fctrl_IP1.bam', 'aligned\u002Fctrl_IP2.bam', 'aligned\u002Fctrl_IP3.bam'), peak_saf)\nip_treat  \u003C- count_matrix(c('aligned\u002Ftreat_IP1.bam', 'aligned\u002Ftreat_IP2.bam', 'aligned\u002Ftreat_IP3.bam'), peak_saf)\nin_ctrl   \u003C- count_matrix(c('aligned\u002Fctrl_Input1.bam', 'aligned\u002Fctrl_Input2.bam', 'aligned\u002Fctrl_Input3.bam'), peak_saf)\nin_treat  \u003C- count_matrix(c('aligned\u002Ftreat_Input1.bam', 'aligned\u002Ftreat_Input2.bam', 'aligned\u002Ftreat_Input3.bam'), peak_saf)\n\nqnb_result \u003C- qnbtest(\n    control_ip    = ip_ctrl,\n    treated_ip    = ip_treat,\n    control_input = in_ctrl,\n    treated_input = in_treat,\n    mode          = 'per-condition'\n)\n\nhead(qnb_result)\nsig \u003C- qnb_result[qnb_result$padj \u003C 0.05 & abs(qnb_result$log2.RR) > 0.5, ]\nnrow(sig)\n```\n\nVerify QNB argument names against `?qnbtest` for the installed version; older tutorials may show different signatures.\n\n## RADAR Reproducibility-Aware Differential\n\n**Goal:** Test differential m6A using a Poisson-NB framework with TMM normalisation and explicit replicate-variance modeling; useful when replicate variability is a known issue.\n\n**Approach:** RADAR's documented workflow is `countReads -> normalizeLibrary -> adjustExprLevel -> filterBins -> diffIP -> reportResult`. CRITICAL: RADAR expects matched BAMs in `bamFolder` named `\u003Csample>.input.bam` and `\u003Csample>.m6A.bam` per replicate; the generic IP \u002F Input naming used elsewhere must be re-conformed or symlinked. `variable()` is set with a data.frame, NOT a bare factor.\n\n```r\nlibrary(RADAR)\n\nradar \u003C- countReads(\n    samplenames  = c('ctrl_rep1', 'ctrl_rep2', 'ctrl_rep3', 'treat_rep1', 'treat_rep2', 'treat_rep3'),\n    gtf          = 'refs\u002Fannotation.gtf',\n    bamFolder    = 'aligned_radar\u002F',\n    modification = 'm6A',\n    strandToKeep = 'opposite',\n    threads      = 8\n)\n\nradar \u003C- normalizeLibrary(radar)\nradar \u003C- adjustExprLevel(radar)\n\nvariable(radar) \u003C- data.frame(group = c('ctrl', 'ctrl', 'ctrl', 'treat', 'treat', 'treat'))\n\nradar \u003C- filterBins(radar, minCountsCutOff = 15)\nradar \u003C- diffIP(radar)\n\nresult \u003C- reportResult(radar, cutoff = 0.1, Beta_cutoff = 0.5)\nsig \u003C- result[result$padj \u003C 0.05 & abs(result$logFC) > 0.5, ]\nnrow(sig)\n```\n\n`reportResult` thresholds (`cutoff` = p-value cutoff; `Beta_cutoff` = effect-size cutoff in beta units) are RADAR-specific; convert to the project's standard reporting thresholds downstream. The `aligned_radar\u002F` directory should contain BAMs with RADAR's expected naming (`\u003Csample>.input.bam`, `\u003Csample>.m6A.bam`).\n\n## Volcano Plot of Differential Peaks\n\n**Goal:** Visualise the differential peak set with effect size on the x-axis and statistical significance on the y-axis; flag peaks passing |log2FC| and FDR thresholds.\n\n**Approach:** Standard ggplot2 volcano with colour-coded significance and threshold lines.\n\n```r\nlibrary(ggplot2)\n\ndiff_table$significance \u003C- with(diff_table,\n    ifelse(padj \u003C 0.05 & abs(log2FC) > 0.5, 'differential', 'not_sig'))\n\nggplot(diff_table, aes(x=log2FC, y=-log10(padj), colour=significance)) +\n    geom_point(alpha=0.5, size=0.8) +\n    geom_vline(xintercept=c(-0.5, 0.5), linetype='dashed') +\n    geom_hline(yintercept=-log10(0.05), linetype='dashed') +\n    scale_colour_manual(values=c(differential='red', not_sig='grey60')) +\n    labs(x='log2 (treat \u002F ctrl) MeRIP enrichment ratio',\n         y='-log10 (FDR)',\n         title='Differential m6A peaks: ctrl vs treat',\n         caption='Per-window IP\u002Finput ratio normalised; not absolute stoichiometry') +\n    theme_minimal()\n```\n\nThe caption is intentional: MeRIP differential reports CHANGES IN ENRICHMENT RATIO, NOT changes in absolute stoichiometry. For stoichiometry claims, cross-validate with GLORI \u002F SAC-seq \u002F m6Anet.\n\n## Per-Method Failure Modes\n\n### Reporting \"hyper-methylation\" without orthogonal calibration\n\n**Trigger:** \"Peak X shows hyper-methylation in treatment\" inferred from MeRIP IP fold-change alone.\n\n**Mechanism:** MeRIP IP fold-change conflates per-molecule methylation stoichiometry, transcript abundance, and IP efficiency variation between libraries. An IP fold-change increase can reflect any or all of these.\n\n**Symptom:** Reported m6A \"hyper-methylation\" tracks RNA-seq expression changes between conditions; reverse-direction effects when properly normalised against input.\n\n**Fix:** Use \"increased \u002F decreased enrichment\" terminology for MeRIP-only studies. Reserve \"hyper- \u002F hypo-methylated\" for studies with absolute quantification orthogonal validation (GLORI, SAC-seq, MAZTER-seq, m6Anet per-read). For high-stakes claims at named loci, run GLORI on a subset of conditions.\n\n### Effect-size threshold absence\n\n**Trigger:** Reporting \"1,500 differential m6A peaks\" with FDR \u003C 0.05 (uncorrected p-value or naive multiple testing) and no effect-size filter.\n\n**Mechanism:** With sufficient sequencing depth, MeRIP-seq has high statistical power to detect very small (~1.1-1.2x) IP-ratio changes that lie within antibody \u002F technical noise. McIntyre 2020 *Sci Rep* 10:6590 showed these changes do not replicate.\n\n**Symptom:** Differential peak set has many peaks with small effect sizes; replication in an independent study recovers \u003C30% of original calls.\n\n**Fix:** Apply effect-size filter (|log2FC| >= 0.5 minimum, often >= 1) AND adjusted p-value (FDR \u003C 0.05) AND replicate-direction concordance. Differential peaks should be reported with effect size, not just p-value. Report effect-size distribution alongside peak count.\n\n### Underpowered N=2 design\n\n**Trigger:** Differential m6A study with N=2 IP and N=2 input per condition.\n\n**Mechanism:** Per McIntyre 2020 and many subsequent benchmarks, MeRIP replicate variance is high; N=2 estimates of dispersion are unreliable; differential calls are unstable.\n\n**Symptom:** Many \"differential\" peaks; volcano plot dense; small fraction replicates in held-out replicate.\n\n**Fix:** Minimum N=3 per condition (per condition per IP\u002Finput arm = 12 BAMs for a 2-condition study); N=4-5 preferred for high-stakes claims. Underpowered studies should report effect-size-only filtered subsets (the most extreme peaks) and acknowledge the noise floor explicitly.\n\n### Batch confounded with condition (antibody lot, IP day)\n\n**Trigger:** Control samples processed in batch 1 with antibody lot A; treatment samples processed in batch 2 with antibody lot B.\n\n**Mechanism:** Anti-m6A antibody lots have batch-to-batch variability in pulldown efficiency and m6A-vs-m6Am cross-reactivity. Pooling cross-batch counts in a differential model attributes batch-effect to condition.\n\n**Symptom:** \"Differential\" peaks at high-abundance transcripts; effect sizes track batch rather than condition; reanalysis with `batch` in the model removes most differential peaks.\n\n**Fix:** Include `antibody_lot` \u002F `batch` \u002F `prep_day` as a fixed effect in a DESeq2 \u002F edgeR model on featureCounts-on-peaks counts; OR run exomePeak2 separately per lot and meta-analyse the per-lot differential peak sets; OR re-design the experiment to avoid lot-condition confounding. exomePeak2's top-level API does NOT accept arbitrary covariates — DESeq2 \u002F edgeR is the route when covariate handling is required.\n\n### exomePeak2 invoked with a fabricated `mode=` argument\n\n**Trigger:** `exomePeak2(..., mode='differential')` OR `exomePeak2(..., mode='diff_peak')` returns \"unused argument\" error.\n\n**Mechanism:** exomePeak2 has NO `mode=` argument. Differential is triggered by populating `bam_treated_ip` and `bam_treated_input` alongside the standard `bam_ip` and `bam_input` (control arm). The `peak_calling_mode` argument is unrelated — it accepts `'exon' | 'full_transcript' | 'whole_genome'` and controls locus scope, not differential-vs-non-differential.\n\n**Fix:** Populate the four BAM-vector arguments (`bam_ip`, `bam_input`, `bam_treated_ip`, `bam_treated_input`); drop any `mode=` reference; consult `?exomePeak2` for the authoritative signature in the installed version.\n\n### Treating peak count matrices like RNA count matrices in edgeR\u002FDESeq2\n\n**Trigger:** Compute featureCounts at peaks, build a count matrix, run edgeR \u002F DESeq2 on the matrix as if peaks were genes.\n\n**Mechanism:** Peak counts reflect both IP enrichment AND transcript abundance. Generic RNA-seq DE on peak counts mixes the two; size-factor normalisation on IP-only counts loses the input-pair information.\n\n**Fix:** edgeR \u002F DESeq2 on peak counts is defensible ONLY for paired symmetric designs where input is modelled as an offset (per-sample size factor on input counts AND per-sample size factor on IP counts, then differential on the ratio). For most uses, exomePeak2 \u002F QNB \u002F RADAR's purpose-built models are more appropriate.\n\n### Counting reads at peaks WITHOUT featureCounts strand-awareness\n\n**Trigger:** `featureCounts(...)` invoked without `strandSpecific=` flag; or with wrong strand setting.\n\n**Mechanism:** Strand-aware counting matters when the protocol is stranded (most modern MeRIP is unstranded; some are reverse-stranded). Wrong strand counts include antisense reads as if they were sense.\n\n**Fix:** Verify protocol strandedness from sequencing-core notes or by inspecting featureCounts summary at a few transcripts. Pass `strandSpecific=0` (unstranded), `1` (forward), or `2` (reverse) explicitly.\n\n## Reconciliation: When Differential Methods Disagree\n\n| Pattern | Likely cause | Action |\n|---------|--------------|--------|\n| exomePeak2 calls a differential peak; QNB does not | Different dispersion estimates; QNB more conservative at low coverage | Trust intersection; report concordant set as high-confidence |\n| Most \"differential\" peaks fall at high-abundance housekeeping transcripts | Expression \u002F IP-efficiency confound | Check log2FC vs input log2FC; if correlated, batch effect or expression-driven |\n| RADAR vs exomePeak2 disagree on direction | Different normalisation (TMM vs internal) | Inspect normalisation diagnostic; choose method aligned with experimental design |\n| Differential peaks anti-correlated with RNA-seq DE | Expression conflated with methylation in the differential model | Re-run with stronger input normalisation; consider per-peak ratio normalisation |\n| Single differential peak survives across all methods | High-confidence call | Orthogonally validate (GLORI \u002F SAC-seq) at the named locus |\n| Differential calls scatter randomly across genome | Underpowered; technical noise dominates | Increase N; apply stricter effect-size filter; report null result honestly |\n| Cross-condition peak overlap \u003C 50% before differential | Conditions are biologically very different; OR antibody lot effect | Inspect cross-replicate concordance; check antibody lot metadata |\n| Volcano shows extreme outliers at low-coverage peaks | Per-peak variance dominated by Poisson sampling | Filter peaks by minimum coverage (>=30 reads in IP AND input) before differential |\n\n## Quantitative Thresholds\n\n| Quantity | Threshold | Source \u002F rationale |\n|----------|-----------|--------------------|\n| Minimum biological replicates per condition | 3 (4-5 preferred) | McIntyre 2020 *Sci Rep* 10:6590 — N=2 routinely under-powered |\n| FDR threshold | 0.05 | Standard convention |\n| Effect-size threshold (|log2FC|) | >= 0.5 minimum; >= 1 stringent | Below 0.5, calls within technical noise floor |\n| Minimum coverage per peak window (each IP AND input) | 30 reads | Standard convention; below this, statistical calls noisy |\n| Replicate-direction concordance | Same direction in >=2\u002FN replicates | Guardrail against single-replicate artifacts |\n| Antibody lot tracking | Mandatory in design matrix when lots differ | Lot confounding is a known false-positive source |\n| MeRIP per-window IP\u002Finput ratio inflation | log2(IP\u002Finput) >= 1 for \"enriched\"; >=2 for \"strongly enriched\" | Convention |\n| Cross-study peak overlap baseline | ~45% median between labs | McIntyre 2020 *Sci Rep* 10:6590 — bounds inter-study reproducibility |\n| Effect size for biological validation | |log2FC| >= 1 AND padj \u003C 0.05 typically used for downstream wet-lab follow-up | Field convention; tighten for low-N studies |\n| GLORI orthogonal validation threshold | Stoichiometry change >= 10% at named site | Liu C 2023 *Nat Biotechnol* 41:355 calibration |\n\n## Common Errors\n\n| Error \u002F symptom | Cause | Solution |\n|-----------------|-------|----------|\n| exomePeak2 `mode='differential'` rejected | exomePeak2 has no `mode=` arg; differential is via `bam_treated_ip` + `bam_treated_input` | Populate the four BAM-vector args; consult `?exomePeak2` |\n| QNB error \"unused argument\" | Function signature changed between versions | Pin commit SHA; verify against installed `?qnbtest` |\n| RADAR install fails | GitHub-only; requires devtools and specific Bioconductor deps | `devtools::install_github('scottzijiezhang\u002FRADAR')`; check Bioconductor requirements |\n| featureCounts returns zero counts | Strand setting wrong; OR peak BED has different chromosome naming than BAM | Verify `strandSpecific=`; reconcile chromosome names |\n| Volcano plot all peaks near origin | Low effect sizes; technical noise dominant | Check input normalisation; increase N; report null result if appropriate |\n| FDR-significant but small effect size | Large dataset finds small differences with high power | Apply effect-size filter; report effect-size distribution |\n| Many differential peaks track expression changes | Input normalisation insufficient | Re-run with stronger input adjustment; OR use ratio-of-ratios |\n| edgeR `estimateDisp()` fails at N=2 | Small-N dispersion estimation collapses | Use QNB instead; designed for this case |\n| MeTDiff install fails | GitHub-only, bundled with MeTPeak | `devtools::install_github('compgenomics\u002FMeTPeak')` |\n| exomePeak2 result has no log2FC \u002F padj columns | `bam_treated_ip` \u002F `bam_treated_input` were not populated; only ran peak calling on the control arm | Pass all four BAM-vector arguments (`bam_ip`, `bam_input`, `bam_treated_ip`, `bam_treated_input`) to trigger differential output |\n| Per-peak boxplot shows huge within-condition variance | High biological noise; OR one replicate is an outlier | Inspect plotCorrelation in merip-preprocessing for outlier; consider exclusion |\n\n## Anticipated Reviewer Pushback\n\n| Pushback | Response |\n|----------|----------|\n| \"How many replicates per condition?\" | N=3 minimum; N=4-5 preferred; rationale McIntyre 2020 |\n| \"What's the effect-size threshold?\" | |log2FC| >= 0.5 minimum; results table reports effect size alongside FDR |\n| \"Was batch \u002F antibody lot controlled for?\" | Yes — `antibody_lot` included as fixed effect in DESeq2 model on featureCounts-on-peaks counts (exomePeak2's top-level API does not accept covariates); per-lot exomePeak2 + meta-analysis when DESeq2 path infeasible; lot-condition confounding assessed |\n| \"Does the differential signal track expression changes?\" | Cross-checked log2FC vs input log2FC per peak; only peaks with strong IP\u002Finput ratio shift reported as differential |\n| \"Was orthogonal validation done?\" | Top hits orthogonally validated via GLORI \u002F m6Anet \u002F per-locus assay |\n| \"Why exomePeak2 over QNB?\" | exomePeak2 default for standard 3-vs-3; QNB used for small-N sensitivity analyses; both reported as concordance check |\n| \"What about absolute stoichiometry?\" | MeRIP differential reports relative enrichment changes; absolute stoichiometry requires GLORI \u002F SAC-seq |\n| \"Was the right normalisation used?\" | Per-window IP\u002Finput ratio normalisation; exomePeak2 internal; alternative TMM (RADAR) compared |\n| \"How does this replicate in independent studies?\" | Cross-study peak overlap reported; differential subset checked against published m6A-Atlas |\n| \"Why not edgeR \u002F DESeq2?\" | Generic RNA-seq DE on peak counts loses IP\u002Finput pairing structure; used only for sensitivity analysis with paired symmetric design |\n\n## References\n\n- Dominissini D, Moshitch-Moshkovitz S, Schwartz S et al (2012) Topology of the human and mouse m6A RNA methylomes revealed by m6A-seq. *Nature* 485(7397):201-206. doi:10.1038\u002Fnature11112\n- Meyer KD, Saletore Y, Zumbo P, Elemento O, Mason CE, Jaffrey SR (2012) Comprehensive analysis of mRNA methylation reveals enrichment in 3' UTRs and near stop codons. *Cell* 149(7):1635-1646. doi:10.1016\u002Fj.cell.2012.05.003\n- Liu L, Zhang SW, Huang Y, Meng J (2017) QNB: differential RNA methylation analysis for count-based small-sample sequencing data with a quad-negative binomial model. *BMC Bioinformatics* 18(1):387. doi:10.1186\u002Fs12859-017-1808-4\n- Cui X, Meng J, Zhang S, Chen Y, Huang Y (2016) A novel algorithm for calling mRNA m6A peaks by modeling biological variances in MeRIP-seq data. *Bioinformatics* 32(12):i378-i385. doi:10.1093\u002Fbioinformatics\u002Fbtw281\n- Liu L, Zhang SW, Gao F et al (2016) DRME: count-based differential RNA methylation analysis at small sample size scenario. *Anal Biochem* 499:15-23. doi:10.1016\u002Fj.ab.2016.01.014\n- Zhang Z, Zhan Q, Eckert M et al (2019) RADAR: differential analysis of MeRIP-seq data with a random effect model. *Genome Biol* 20(1):294. doi:10.1186\u002Fs13059-019-1915-9\n- Meng J, Lu Z, Liu H et al (2014) A protocol for RNA methylation differential analysis with MeRIP-Seq data and exomePeak R\u002FBioconductor package. *Methods* 69(3):274-281. doi:10.1016\u002Fj.ymeth.2014.06.008\n- Liu J, Zhang Z, Meng J et al (2022) exomePeak2: a peak calling and differential analysis tool for MeRIP-Seq with bias awareness. *NAR Genom Bioinform* 4(3):lqac046. doi:10.1093\u002Fnargab\u002Flqac046\n- McIntyre ABR, Gokhale NS, Cerchietti L, Jaffrey SR, Horner SM, Mason CE (2020) Limits in the detection of m6A changes using MeRIP\u002Fm6A-seq. *Sci Rep* 10(1):6590. doi:10.1038\u002Fs41598-020-63355-3\n- Liu C, Sun H, Yi Y et al (2023) Absolute quantification of single-base m6A methylation in the mammalian transcriptome using GLORI. *Nat Biotechnol* 41(3):355-366. doi:10.1038\u002Fs41587-022-01487-9\n- Hu L, Liu S, Peng Y et al (2022) m6A RNA modifications are measured at single-base resolution across the mammalian transcriptome. *Nat Biotechnol* 40(8):1210-1219. doi:10.1038\u002Fs41587-022-01243-z\n- Garcia-Campos MA, Edelheit S, Toth U et al (2019) Deciphering the m6A code via antibody-independent quantitative profiling. *Cell* 178(3):731-747.e16. doi:10.1016\u002Fj.cell.2019.06.013\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- 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- Liao Y, Smyth GK, Shi W (2014) featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. *Bioinformatics* 30(7):923-930. doi:10.1093\u002Fbioinformatics\u002Fbtt656\n\n## Related Skills\n\n- merip-preprocessing - Upstream IP\u002Finput BAM preparation; design-matrix metadata (antibody lot, batch) originates here\n- m6a-peak-calling - Peak calling step that produces input to differential analysis\n- m6anet-analysis - Orthogonal ONT-direct-RNA validation for stoichiometry claims at high-stakes loci\n- modification-visualization - Volcano \u002F MA \u002F per-peak boxplot rendering of differential results\n- differential-expression\u002Fdeseq2-basics - Canonical DE design philosophy; m6a-differential defers to this for general design-matrix patterns\n- differential-expression\u002Fde-results - Post-DE interpretation, ranking, gene-list extraction\n- differential-expression\u002Fedger-basics - edgeR fundamentals for the paired-symmetric sensitivity case\n- chip-seq\u002Fdifferential-binding - Closest sibling for IP-vs-input differential binding (general framework)\n- rna-quantification\u002Ffeaturecounts-counting - Peak count matrix construction\n- data-visualization\u002Fvolcano-and-ma-plots - Volcano + MA plot recipes\n- data-visualization\u002Fmultipanel-figures - Figure assembly\n- pathway-analysis\u002Fgo-enrichment - GO enrichment on differential-peak-bearing gene lists\n- workflows\u002Frnaseq-to-de - End-to-end pipeline orchestration patterns\n","## Version Compatibility\n\nReference examples tested with: exomePeak2 1.14+ (Bioconductor 3.18+), QNB 1.1.11 (GitHub `lzcyzm\u002FQNB`), MeTDiff (GitHub, bundled with MeTPeak), RADAR 0.2.4+ (GitHub `scottzijiezhang\u002FRADAR`), DESeq2 1.42+, edgeR 4.0+, GenomicFeatures 1.54+, ggplot2 3.5+, GenomicAlignments 1.38+, Rsubread 2.16+.\n\nBefore using code patterns, verify installed versions match. If versions differ:\n- R: `packageVersion('exomePeak2')` then `?exomePeak2` to verify parameters\n- R: `packageVersion('QNB')` then `?qnbtest` to confirm argument signature\n\nIf R throws `unused argument` or `argument is missing`, the API moved between Bioconductor minor releases; consult `?exomePeak2` directly. QNB function signature has been stable since 2017 but the GitHub source has occasional changes; pin the commit SHA.\n\nexomePeak2's differential interface is NOT a `mode=` argument — populate the `bam_treated_ip=` and `bam_treated_input=` arguments alongside the standard `bam_ip=` \u002F `bam_input=` (control arm) to trigger paired differential calling. `peak_calling_mode` is a separate argument controlling locus scope (`exon` | `full_transcript` | `whole_genome`). Verify against `?exomePeak2` and Bioconductor 3.20+ release notes. QNB is on GitHub only; install via `devtools::install_github('lzcyzm\u002FQNB')`. RADAR is on GitHub only; install via `devtools::install_github('scottzijiezhang\u002FRADAR')`; its differential workflow is `countReads -> normalizeLibrary -> adjustExprLevel -> filterBins -> diffIP -> reportResult`.\n\n# Differential m6A Analysis\n\n**\"Compare m6A methylation between my conditions\"** -> Quantify how much each m6A peak's IP\u002Finput enrichment shifts between conditions, after normalising for transcript-abundance changes (which all show up in input) and for IP-efficiency drift (which the design matrix and within-run replicates control for). Then apply effect-size filtering to distinguish real biology from the technical noise floor that all MeRIP differential methods inherit (McIntyre 2020 *Sci Rep* 10:6590: between-study m6A peak overlap is ~45% median; differential calls within that noise envelope routinely fail to replicate). Critically: a higher MeRIP signal in condition A vs B can mean (1) more transcripts of the peak-bearing gene, (2) more methylation per transcript, OR (3) higher IP efficiency in batch A — distinguishing requires careful normalisation or an orthogonal absolute-stoichiometry method.\n\n- R: `exomePeak2::exomePeak2(bam_ip, bam_input, bam_treated_ip, bam_treated_input, ...)` -- integrated peak + differential GLM (modern default)\n- R: `QNB::qnbtest(control_ip, treated_ip, control_input, treated_input)` -- beta-binomial for small N (Liu 2017 *BMC Bioinformatics* 18:387)\n- R: `RADAR::diffIP()` then `reportResult()` -- Poisson-NB on peak windows with TMM normalisation (Zhang 2019 *Genome Biol* 20:294)\n- R: `MeTDiff::metdiff()` -- HMM-based differential paired with MeTPeak\n- R: edgeR \u002F DESeq2 on featureCounts-on-peaks matrix -- defensible for paired symmetric designs with strong input normalisation; also the route for arbitrary batch \u002F lot covariates exomePeak2's API does not accept\n\n## The Single Most Important Modern Insight -- IP fold-change between conditions conflates stoichiometry change with expression change and IP efficiency drift\n\nA higher m6A peak signal in condition A vs B can mean ANY of: (1) more transcripts of the peak-bearing gene (expression up; the input increases proportionally, so the ratio should not change — but residual normalisation noise leaks through), (2) more methylation per transcript (stoichiometry up; the real biology of interest), (3) higher IP efficiency in batch A (technical; antibody lot, IP day, RNA prep). Differential MeRIP WITHOUT per-window input-normalisation OR an orthogonal stoichiometry-aware method (GLORI Liu 2023 *Nat Biotechnol* 41:355; SAC-seq Hu 2022 *Nat Biotechnol* 40:1210; MAZTER-seq Garcia-Campos 2019 *Cell* 178:731; m6Anet per-read modification rate) cannot separate the three. exomePeak2 differential mode, QNB, RADAR, and MeTDiff all implement per-window IP\u002Finput ratio modelling, but each has different default normalisations and the choice matters. Equally critical: McIntyre 2020 *Sci Rep* 10:6590 showed that \"differential\" m6A peaks from MeRIP-seq routinely do not replicate between independent studies in nominally identical conditions. The empirical noise floor is high; effect-size filtering (|log2FC| >= 0.5 minimum, often >= 1) AND replicate-direction concordance (the change is consistent in direction across replicates) AND minimum N=3 per condition are needed for differential calls to survive replication. For any absolute stoichiometry claim (\"this peak is 80% methylated in tumour vs 20% in normal\"), require an orthogonal stoichiometry method, not MeRIP alone.\n\n## Algorithmic Taxonomy\n\n| Tool \u002F mode | Mechanism | Inputs | Output | Strength | Fails when |\n|-------------|-----------|--------|--------|----------|------------|\n| exomePeak2 differential (Liu 2022) | Transcript-windowed Poisson GLM with GC-bias correction; integrated peak + differential | control IP\u002Finput BAM vectors + treated IP\u002Finput BAM vectors + TxDb + BSgenome | Differential peaks with log2FC + FDR per peak | Most use cases; integrates with peak calling; modern default | Small-N (\u003C=2) overdispersion poorly estimated; top-level API does NOT accept arbitrary covariates (use DESeq2\u002FedgeR for batch-aware designs) |\n| QNB (Liu 2017 *BMC Bioinformatics* 18:387) | Quad-negative-binomial joint model of IP, input, condition | per-peak count matrices (4 matrices: ip1, ip2, input1, input2) | Differential peaks with p-value + log2 RR | Designed for small N (2-3 per group); handles overdispersion explicitly | Requires pre-computed count matrices; not integrated with peak calling |\n| MeTDiff (bundled with MeTPeak) | HMM + Beta-binomial differential paired with MeTPeak | paired IP\u002Finput BAM + GTF + condition factor | Differential peaks per window | Pairs naturally with MeTPeak output; HMM smoothing helps low-coverage | GitHub-only; less benchmarked than exomePeak2 \u002F QNB |\n| RADAR (Zhang 2019 *Genome Biol* 20:294) | Poisson-NB with TMM normalisation; reproducibility-aware | paired IP\u002Finput BAM + condition factor | Differential peaks with logFC + FDR | Explicit replicate-variance modeling; reproducibility-aware framework | GitHub-only; slower than exomePeak2 |\n| DRME (Liu 2016 *Anal Biochem* 499:15) | Count-based small-sample alternative | per-peak count matrices | Differential peaks | Sister to QNB from same group; small-N alternative | Less benchmarked than QNB \u002F exomePeak2 |\n| edgeR \u002F DESeq2 on peak counts | Generic RNA-seq differential framework applied to featureCounts-on-peaks | peak count matrix + sample sheet | Differential peaks with log2FC + FDR | Familiar; flexible designs; well-tested in RNA-seq | Treats peak counts as RNA counts; loses IP\u002Finput pairing structure; defensible only for paired symmetric designs with strong input normalisation |\n| Ratio-of-ratios (heuristic) | Compute per-peak log2 (IP_A \u002F Input_A) - log2 (IP_B \u002F Input_B) per sample, then t-test | per-peak count matrix | per-peak t-test | Transparent; no model assumptions | No multiple-testing correction; ignores overdispersion; not recommended for primary analysis |\n\n## Decision Tree by Scenario\n\n| Scenario | Recommended | Why wrong choices fail |\n|----------|-------------|------------------------|\n| Standard 3-vs-3 paired-design MeRIP differential | exomePeak2 with `bam_ip` (ctrl) + `bam_treated_ip` (treat) and matching inputs | QNB usable but designed for smaller N; edgeR\u002FDESeq2 loses IP\u002Finput pairing |\n| Very small N (2 vs 2) | QNB (designed for small-sample overdispersion); supplement with exomePeak2 if possible | edgeR \u002F DESeq2 dispersion estimation collapses; exomePeak2 GLM also struggles at N=2 |\n| Paired design (patient as blocking factor) | QNB per-pair then aggregate; OR featureCounts-on-peaks -> DESeq2 with `~patient + condition` design; exomePeak2 cannot encode patient blocking via its top-level API | Unpaired analysis inflates within-group variance |\n| Interaction design (genotype × treatment) | featureCounts-on-peaks -> DESeq2 \u002F edgeR with interaction term; exomePeak2 top-level API is two-group only | QNB pairwise only; build interaction model from pairwise contrasts manually |\n| Batch confounding (antibody lot, sequencing run, IP day) | Include batch as fixed effect in DESeq2 \u002F edgeR model on featureCounts-on-peaks matrix; OR run exomePeak2 per-batch and meta-analyse | Pooling cross-batch counts without batch term attributes lot-effect to condition |\n| Time-course differential | featureCounts-on-peaks -> DESeq2 \u002F limma with time as numeric covariate; OR pairwise time-point exomePeak2 contrasts | Naive group-vs-group ignores time structure |\n| Stoichiometry claims (not just enrichment) | NOT MeRIP differential -- orthogonal GLORI \u002F SAC-seq \u002F m6Anet per-read | MeRIP IP fold-change is relative; cannot give per-molecule stoichiometry |\n| Cross-batch differential (different antibody lots) | featureCounts-on-peaks -> DESeq2 \u002F edgeR with `lot` in design; OR run exomePeak2 per-lot then meta-analyse; ideally avoid confounding lot with condition at the experimental-design stage | Lot effect inflates false positives; exomePeak2 top-level API cannot encode lot |\n| Validation of differential calls | Run >=2 differential methods; require concordant direction across replicates AND |log2FC| > 0.5 minimum (>= 1 stringent); orthogonal validation at top hits | Single-method differential calls within technical noise floor (per McIntyre 2020) |\n| Visualising differential peaks | Volcano plot with |log2FC| + FDR thresholds; MA plot to inspect normalisation; per-peak boxplot for top hits | Single number summaries hide stoichiometry vs expression confound |\n| Wanting to test a single gene \u002F locus | Targeted: per-peak boxplot across replicates with condition factor; manual t-test or Wilcoxon at high-coverage peak | Whole-transcriptome differential testing wastes multiple-testing budget for single-locus questions |\n\nMethodology evolves; before any high-stakes differential analysis, web-search \"exomePeak2 differential mode Bioconductor 3.20\" and \"MeRIP differential benchmark McIntyre\" for current consensus parameters.\n\n## exomePeak2 Differential Workflow\n\n**Goal:** Identify m6A peaks that differ in methylation level between conditions, controlling for transcript-abundance differences (via input normalisation) and GC bias (via internal correction), with an integrated peak-calling + differential pipeline.\n\n**Approach:** Build TxDb from the matched GTF; pass control IP\u002Finput BAM vectors via `bam_ip` and `bam_input` AND treatment IP\u002Finput BAM vectors via `bam_treated_ip` and `bam_treated_input` — populating the treated arms triggers differential mode (there is no separate `mode=` argument). Output is per-peak log2FC + FDR.\n\n```r\nlibrary(exomePeak2)\nlibrary(GenomicFeatures)\nlibrary(BSgenome.Hsapiens.UCSC.hg38)\n\ntxdb \u003C- makeTxDbFromGFF('refs\u002Fannotation.gtf', format='gtf')\n\nctrl_ip      \u003C- c('aligned\u002Fctrl_IP1.bam', 'aligned\u002Fctrl_IP2.bam', 'aligned\u002Fctrl_IP3.bam')\nctrl_input   \u003C- c('aligned\u002Fctrl_Input1.bam', 'aligned\u002Fctrl_Input2.bam', 'aligned\u002Fctrl_Input3.bam')\ntreat_ip     \u003C- c('aligned\u002Ftreat_IP1.bam', 'aligned\u002Ftreat_IP2.bam', 'aligned\u002Ftreat_IP3.bam')\ntreat_input  \u003C- c('aligned\u002Ftreat_Input1.bam', 'aligned\u002Ftreat_Input2.bam', 'aligned\u002Ftreat_Input3.bam')\n\nresult \u003C- exomePeak2(\n    bam_ip             = ctrl_ip,\n    bam_input          = ctrl_input,\n    bam_treated_ip     = treat_ip,\n    bam_treated_input  = treat_input,\n    txdb               = txdb,\n    genome             = BSgenome.Hsapiens.UCSC.hg38,\n    paired_end         = TRUE,\n    library_type       = 'unstranded',\n    peak_calling_mode  = 'exon',\n    save_dir           = 'exomepeak2_diff_output',\n    experiment_name    = 'ctrl_vs_treat'\n)\n\ndiff_table \u003C- as.data.frame(result)\nnrow(diff_table)\nhead(diff_table[, c('seqnames', 'start', 'end', 'log2FC', 'pvalue', 'padj')])\n```\n\n`peak_calling_mode` accepts `'exon'` (transcript-aware, default), `'full_transcript'`, or `'whole_genome'`; the meaning is locus scope, NOT differential-vs-non-differential. For arbitrary covariate adjustment (batch, antibody lot, patient blocking), the exomePeak2 top-level API is insufficient — move counts into DESeq2 \u002F edgeR via the featureCounts-on-peaks route below.\n\n## QNB Beta-Binomial for Small-N Designs\n\n**Goal:** Test differential m6A at pre-called peaks using a quad-negative-binomial model that handles small-N overdispersion better than generic GLM frameworks.\n\n**Approach:** Count reads in each IP and Input BAM at each peak using featureCounts or summarizeOverlaps; pass the four count matrices (ip1, ip2, input1, input2 — ip\u002Finput per group) to `qnbtest()`.\n\n```r\nlibrary(QNB)\nlibrary(Rsubread)\nlibrary(rtracklayer)\n\npeaks \u003C- import('exomepeak2_output\u002Fm6a_run1\u002Fpeaks.bed')\npeak_saf \u003C- data.frame(\n    GeneID = paste0('peak_', seq_along(peaks)),\n    Chr    = as.character(seqnames(peaks)),\n    Start  = start(peaks),\n    End    = end(peaks),\n    Strand = as.character(strand(peaks))\n)\n\ncount_matrix \u003C- function(bam_paths, peak_saf) {\n    fc \u003C- featureCounts(\n        files       = bam_paths,\n        annot.ext   = peak_saf,\n        isPairedEnd = TRUE,\n        nthreads    = 8,\n        allowMultiOverlap = TRUE\n    )\n    fc$counts\n}\n\nip_ctrl   \u003C- count_matrix(c('aligned\u002Fctrl_IP1.bam', 'aligned\u002Fctrl_IP2.bam', 'aligned\u002Fctrl_IP3.bam'), peak_saf)\nip_treat  \u003C- count_matrix(c('aligned\u002Ftreat_IP1.bam', 'aligned\u002Ftreat_IP2.bam', 'aligned\u002Ftreat_IP3.bam'), peak_saf)\nin_ctrl   \u003C- count_matrix(c('aligned\u002Fctrl_Input1.bam', 'aligned\u002Fctrl_Input2.bam', 'aligned\u002Fctrl_Input3.bam'), peak_saf)\nin_treat  \u003C- count_matrix(c('aligned\u002Ftreat_Input1.bam', 'aligned\u002Ftreat_Input2.bam', 'aligned\u002Ftreat_Input3.bam'), peak_saf)\n\nqnb_result \u003C- qnbtest(\n    control_ip    = ip_ctrl,\n    treated_ip    = ip_treat,\n    control_input = in_ctrl,\n    treated_input = in_treat,\n    mode          = 'per-condition'\n)\n\nhead(qnb_result)\nsig \u003C- qnb_result[qnb_result$padj \u003C 0.05 & abs(qnb_result$log2.RR) > 0.5, ]\nnrow(sig)\n```\n\nVerify QNB argument names against `?qnbtest` for the installed version; older tutorials may show different signatures.\n\n## RADAR Reproducibility-Aware Differential\n\n**Goal:** Test differential m6A using a Poisson-NB framework with TMM normalisation and explicit replicate-variance modeling; useful when replicate variability is a known issue.\n\n**Approach:** RADAR's documented workflow is `countReads -> normalizeLibrary -> adjustExprLevel -> filterBins -> diffIP -> reportResult`. CRITICAL: RADAR expects matched BAMs in `bamFolder` named `\u003Csample>.input.bam` and `\u003Csample>.m6A.bam` per replicate; the generic IP \u002F Input naming used elsewhere must be re-conformed or symlinked. `variable()` is set with a data.frame, NOT a bare factor.\n\n```r\nlibrary(RADAR)\n\nradar \u003C- countReads(\n    samplenames  = c('ctrl_rep1', 'ctrl_rep2', 'ctrl_rep3', 'treat_rep1', 'treat_rep2', 'treat_rep3'),\n    gtf          = 'refs\u002Fannotation.gtf',\n    bamFolder    = 'aligned_radar\u002F',\n    modification = 'm6A',\n    strandToKeep = 'opposite',\n    threads      = 8\n)\n\nradar \u003C- normalizeLibrary(radar)\nradar \u003C- adjustExprLevel(radar)\n\nvariable(radar) \u003C- data.frame(group = c('ctrl', 'ctrl', 'ctrl', 'treat', 'treat', 'treat'))\n\nradar \u003C- filterBins(radar, minCountsCutOff = 15)\nradar \u003C- diffIP(radar)\n\nresult \u003C- reportResult(radar, cutoff = 0.1, Beta_cutoff = 0.5)\nsig \u003C- result[result$padj \u003C 0.05 & abs(result$logFC) > 0.5, ]\nnrow(sig)\n```\n\n`reportResult` thresholds (`cutoff` = p-value cutoff; `Beta_cutoff` = effect-size cutoff in beta units) are RADAR-specific; convert to the project's standard reporting thresholds downstream. The `aligned_radar\u002F` directory should contain BAMs with RADAR's expected naming (`\u003Csample>.input.bam`, `\u003Csample>.m6A.bam`).\n\n## Volcano Plot of Differential Peaks\n\n**Goal:** Visualise the differential peak set with effect size on the x-axis and statistical significance on the y-axis; flag peaks passing |log2FC| and FDR thresholds.\n\n**Approach:** Standard ggplot2 volcano with colour-coded significance and threshold lines.\n\n```r\nlibrary(ggplot2)\n\ndiff_table$significance \u003C- with(diff_table,\n    ifelse(padj \u003C 0.05 & abs(log2FC) > 0.5, 'differential', 'not_sig'))\n\nggplot(diff_table, aes(x=log2FC, y=-log10(padj), colour=significance)) +\n    geom_point(alpha=0.5, size=0.8) +\n    geom_vline(xintercept=c(-0.5, 0.5), linetype='dashed') +\n    geom_hline(yintercept=-log10(0.05), linetype='dashed') +\n    scale_colour_manual(values=c(differential='red', not_sig='grey60')) +\n    labs(x='log2 (treat \u002F ctrl) MeRIP enrichment ratio',\n         y='-log10 (FDR)',\n         title='Differential m6A peaks: ctrl vs treat',\n         caption='Per-window IP\u002Finput ratio normalised; not absolute stoichiometry') +\n    theme_minimal()\n```\n\nThe caption is intentional: MeRIP differential reports CHANGES IN ENRICHMENT RATIO, NOT changes in absolute stoichiometry. For stoichiometry claims, cross-validate with GLORI \u002F SAC-seq \u002F m6Anet.\n\n## Per-Method Failure Modes\n\n### Reporting \"hyper-methylation\" without orthogonal calibration\n\n**Trigger:** \"Peak X shows hyper-methylation in treatment\" inferred from MeRIP IP fold-change alone.\n\n**Mechanism:** MeRIP IP fold-change conflates per-molecule methylation stoichiometry, transcript abundance, and IP efficiency variation between libraries. An IP fold-change increase can reflect any or all of these.\n\n**Symptom:** Reported m6A \"hyper-methylation\" tracks RNA-seq expression changes between conditions; reverse-direction effects when properly normalised against input.\n\n**Fix:** Use \"increased \u002F decreased enrichment\" terminology for MeRIP-only studies. Reserve \"hyper- \u002F hypo-methylated\" for studies with absolute quantification orthogonal validation (GLORI, SAC-seq, MAZTER-seq, m6Anet per-read). For high-stakes claims at named loci, run GLORI on a subset of conditions.\n\n### Effect-size threshold absence\n\n**Trigger:** Reporting \"1,500 differential m6A peaks\" with FDR \u003C 0.05 (uncorrected p-value or naive multiple testing) and no effect-size filter.\n\n**Mechanism:** With sufficient sequencing depth, MeRIP-seq has high statistical power to detect very small (~1.1-1.2x) IP-ratio changes that lie within antibody \u002F technical noise. McIntyre 2020 *Sci Rep* 10:6590 showed these changes do not replicate.\n\n**Symptom:** Differential peak set has many peaks with small effect sizes; replication in an independent study recovers \u003C30% of original calls.\n\n**Fix:** Apply effect-size filter (|log2FC| >= 0.5 minimum, often >= 1) AND adjusted p-value (FDR \u003C 0.05) AND replicate-direction concordance. Differential peaks should be reported with effect size, not just p-value. Report effect-size distribution alongside peak count.\n\n### Underpowered N=2 design\n\n**Trigger:** Differential m6A study with N=2 IP and N=2 input per condition.\n\n**Mechanism:** Per McIntyre 2020 and many subsequent benchmarks, MeRIP replicate variance is high; N=2 estimates of dispersion are unreliable; differential calls are unstable.\n\n**Symptom:** Many \"differential\" peaks; volcano plot dense; small fraction replicates in held-out replicate.\n\n**Fix:** Minimum N=3 per condition (per condition per IP\u002Finput arm = 12 BAMs for a 2-condition study); N=4-5 preferred for high-stakes claims. Underpowered studies should report effect-size-only filtered subsets (the most extreme peaks) and acknowledge the noise floor explicitly.\n\n### Batch confounded with condition (antibody lot, IP day)\n\n**Trigger:** Control samples processed in batch 1 with antibody lot A; treatment samples processed in batch 2 with antibody lot B.\n\n**Mechanism:** Anti-m6A antibody lots have batch-to-batch variability in pulldown efficiency and m6A-vs-m6Am cross-reactivity. Pooling cross-batch counts in a differential model attributes batch-effect to condition.\n\n**Symptom:** \"Differential\" peaks at high-abundance transcripts; effect sizes track batch rather than condition; reanalysis with `batch` in the model removes most differential peaks.\n\n**Fix:** Include `antibody_lot` \u002F `batch` \u002F `prep_day` as a fixed effect in a DESeq2 \u002F edgeR model on featureCounts-on-peaks counts; OR run exomePeak2 separately per lot and meta-analyse the per-lot differential peak sets; OR re-design the experiment to avoid lot-condition confounding. exomePeak2's top-level API does NOT accept arbitrary covariates — DESeq2 \u002F edgeR is the route when covariate handling is required.\n\n### exomePeak2 invoked with a fabricated `mode=` argument\n\n**Trigger:** `exomePeak2(..., mode='differential')` OR `exomePeak2(..., mode='diff_peak')` returns \"unused argument\" error.\n\n**Mechanism:** exomePeak2 has NO `mode=` argument. Differential is triggered by populating `bam_treated_ip` and `bam_treated_input` alongside the standard `bam_ip` and `bam_input` (control arm). The `peak_calling_mode` argument is unrelated — it accepts `'exon' | 'full_transcript' | 'whole_genome'` and controls locus scope, not differential-vs-non-differential.\n\n**Fix:** Populate the four BAM-vector arguments (`bam_ip`, `bam_input`, `bam_treated_ip`, `bam_treated_input`); drop any `mode=` reference; consult `?exomePeak2` for the authoritative signature in the installed version.\n\n### Treating peak count matrices like RNA count matrices in edgeR\u002FDESeq2\n\n**Trigger:** Compute featureCounts at peaks, build a count matrix, run edgeR \u002F DESeq2 on the matrix as if peaks were genes.\n\n**Mechanism:** Peak counts reflect both IP enrichment AND transcript abundance. Generic RNA-seq DE on peak counts mixes the two; size-factor normalisation on IP-only counts loses the input-pair information.\n\n**Fix:** edgeR \u002F DESeq2 on peak counts is defensible ONLY for paired symmetric designs where input is modelled as an offset (per-sample size factor on input counts AND per-sample size factor on IP counts, then differential on the ratio). For most uses, exomePeak2 \u002F QNB \u002F RADAR's purpose-built models are more appropriate.\n\n### Counting reads at peaks WITHOUT featureCounts strand-awareness\n\n**Trigger:** `featureCounts(...)` invoked without `strandSpecific=` flag; or with wrong strand setting.\n\n**Mechanism:** Strand-aware counting matters when the protocol is stranded (most modern MeRIP is unstranded; some are reverse-stranded). Wrong strand counts include antisense reads as if they were sense.\n\n**Fix:** Verify protocol strandedness from sequencing-core notes or by inspecting featureCounts summary at a few transcripts. Pass `strandSpecific=0` (unstranded), `1` (forward), or `2` (reverse) explicitly.\n\n## Reconciliation: When Differential Methods Disagree\n\n| Pattern | Likely cause | Action |\n|---------|--------------|--------|\n| exomePeak2 calls a differential peak; QNB does not | Different dispersion estimates; QNB more conservative at low coverage | Trust intersection; report concordant set as high-confidence |\n| Most \"differential\" peaks fall at high-abundance housekeeping transcripts | Expression \u002F IP-efficiency confound | Check log2FC vs input log2FC; if correlated, batch effect or expression-driven |\n| RADAR vs exomePeak2 disagree on direction | Different normalisation (TMM vs internal) | Inspect normalisation diagnostic; choose method aligned with experimental design |\n| Differential peaks anti-correlated with RNA-seq DE | Expression conflated with methylation in the differential model | Re-run with stronger input normalisation; consider per-peak ratio normalisation |\n| Single differential peak survives across all methods | High-confidence call | Orthogonally validate (GLORI \u002F SAC-seq) at the named locus |\n| Differential calls scatter randomly across genome | Underpowered; technical noise dominates | Increase N; apply stricter effect-size filter; report null result honestly |\n| Cross-condition peak overlap \u003C 50% before differential | Conditions are biologically very different; OR antibody lot effect | Inspect cross-replicate concordance; check antibody lot metadata |\n| Volcano shows extreme outliers at low-coverage peaks | Per-peak variance dominated by Poisson sampling | Filter peaks by minimum coverage (>=30 reads in IP AND input) before differential |\n\n## Quantitative Thresholds\n\n| Quantity | Threshold | Source \u002F rationale |\n|----------|-----------|--------------------|\n| Minimum biological replicates per condition | 3 (4-5 preferred) | McIntyre 2020 *Sci Rep* 10:6590 — N=2 routinely under-powered |\n| FDR threshold | 0.05 | Standard convention |\n| Effect-size threshold (|log2FC|) | >= 0.5 minimum; >= 1 stringent | Below 0.5, calls within technical noise floor |\n| Minimum coverage per peak window (each IP AND input) | 30 reads | Standard convention; below this, statistical calls noisy |\n| Replicate-direction concordance | Same direction in >=2\u002FN replicates | Guardrail against single-replicate artifacts |\n| Antibody lot tracking | Mandatory in design matrix when lots differ | Lot confounding is a known false-positive source |\n| MeRIP per-window IP\u002Finput ratio inflation | log2(IP\u002Finput) >= 1 for \"enriched\"; >=2 for \"strongly enriched\" | Convention |\n| Cross-study peak overlap baseline | ~45% median between labs | McIntyre 2020 *Sci Rep* 10:6590 — bounds inter-study reproducibility |\n| Effect size for biological validation | |log2FC| >= 1 AND padj \u003C 0.05 typically used for downstream wet-lab follow-up | Field convention; tighten for low-N studies |\n| GLORI orthogonal validation threshold | Stoichiometry change >= 10% at named site | Liu C 2023 *Nat Biotechnol* 41:355 calibration |\n\n## Common Errors\n\n| Error \u002F symptom | Cause | Solution |\n|-----------------|-------|----------|\n| exomePeak2 `mode='differential'` rejected | exomePeak2 has no `mode=` arg; differential is via `bam_treated_ip` + `bam_treated_input` | Populate the four BAM-vector args; consult `?exomePeak2` |\n| QNB error \"unused argument\" | Function signature changed between versions | Pin commit SHA; verify against installed `?qnbtest` |\n| RADAR install fails | GitHub-only; requires devtools and specific Bioconductor deps | `devtools::install_github('scottzijiezhang\u002FRADAR')`; check Bioconductor requirements |\n| featureCounts returns zero counts | Strand setting wrong; OR peak BED has different chromosome naming than BAM | Verify `strandSpecific=`; reconcile chromosome names |\n| Volcano plot all peaks near origin | Low effect sizes; technical noise dominant | Check input normalisation; increase N; report null result if appropriate |\n| FDR-significant but small effect size | Large dataset finds small differences with high power | Apply effect-size filter; report effect-size distribution |\n| Many differential peaks track expression changes | Input normalisation insufficient | Re-run with stronger input adjustment; OR use ratio-of-ratios |\n| edgeR `estimateDisp()` fails at N=2 | Small-N dispersion estimation collapses | Use QNB instead; designed for this case |\n| MeTDiff install fails | GitHub-only, bundled with MeTPeak | `devtools::install_github('compgenomics\u002FMeTPeak')` |\n| exomePeak2 result has no log2FC \u002F padj columns | `bam_treated_ip` \u002F `bam_treated_input` were not populated; only ran peak calling on the control arm | Pass all four BAM-vector arguments (`bam_ip`, `bam_input`, `bam_treated_ip`, `bam_treated_input`) to trigger differential output |\n| Per-peak boxplot shows huge within-condition variance | High biological noise; OR one replicate is an outlier | Inspect plotCorrelation in merip-preprocessing for outlier; consider exclusion |\n\n## Anticipated Reviewer Pushback\n\n| Pushback | Response |\n|----------|----------|\n| \"How many replicates per condition?\" | N=3 minimum; N=4-5 preferred; rationale McIntyre 2020 |\n| \"What's the effect-size threshold?\" | |log2FC| >= 0.5 minimum; results table reports effect size alongside FDR |\n| \"Was batch \u002F antibody lot controlled for?\" | Yes — `antibody_lot` included as fixed effect in DESeq2 model on featureCounts-on-peaks counts (exomePeak2's top-level API does not accept covariates); per-lot exomePeak2 + meta-analysis when DESeq2 path infeasible; lot-condition confounding assessed |\n| \"Does the differential signal track expression changes?\" | Cross-checked log2FC vs input log2FC per peak; only peaks with strong IP\u002Finput ratio shift reported as differential |\n| \"Was orthogonal validation done?\" | Top hits orthogonally validated via GLORI \u002F m6Anet \u002F per-locus assay |\n| \"Why exomePeak2 over QNB?\" | exomePeak2 default for standard 3-vs-3; QNB used for small-N sensitivity analyses; both reported as concordance check |\n| \"What about absolute stoichiometry?\" | MeRIP differential reports relative enrichment changes; absolute stoichiometry requires GLORI \u002F SAC-seq |\n| \"Was the right normalisation used?\" | Per-window IP\u002Finput ratio normalisation; exomePeak2 internal; alternative TMM (RADAR) compared |\n| \"How does this replicate in independent studies?\" | Cross-study peak overlap reported; differential subset checked against published m6A-Atlas |\n| \"Why not edgeR \u002F DESeq2?\" | Generic RNA-seq DE on peak counts loses IP\u002Finput pairing structure; used only for sensitivity analysis with paired symmetric design |\n\n## References\n\n- Dominissini D, Moshitch-Moshkovitz S, Schwartz S et al (2012) Topology of the human and mouse m6A RNA methylomes revealed by m6A-seq. *Nature* 485(7397):201-206. doi:10.1038\u002Fnature11112\n- Meyer KD, Saletore Y, Zumbo P, Elemento O, Mason CE, Jaffrey SR (2012) Comprehensive analysis of mRNA methylation reveals enrichment in 3' UTRs and near stop codons. *Cell* 149(7):1635-1646. doi:10.1016\u002Fj.cell.2012.05.003\n- Liu L, Zhang SW, Huang Y, Meng J (2017) QNB: differential RNA methylation analysis for count-based small-sample sequencing data with a quad-negative binomial model. *BMC Bioinformatics* 18(1):387. doi:10.1186\u002Fs12859-017-1808-4\n- Cui X, Meng J, Zhang S, Chen Y, Huang Y (2016) A novel algorithm for calling mRNA m6A peaks by modeling biological variances in MeRIP-seq data. *Bioinformatics* 32(12):i378-i385. doi:10.1093\u002Fbioinformatics\u002Fbtw281\n- Liu L, Zhang SW, Gao F et al (2016) DRME: count-based differential RNA methylation analysis at small sample size scenario. *Anal Biochem* 499:15-23. doi:10.1016\u002Fj.ab.2016.01.014\n- Zhang Z, Zhan Q, Eckert M et al (2019) RADAR: differential analysis of MeRIP-seq data with a random effect model. *Genome Biol* 20(1):294. doi:10.1186\u002Fs13059-019-1915-9\n- Meng J, Lu Z, Liu H et al (2014) A protocol for RNA methylation differential analysis with MeRIP-Seq data and exomePeak R\u002FBioconductor package. *Methods* 69(3):274-281. doi:10.1016\u002Fj.ymeth.2014.06.008\n- Liu J, Zhang Z, Meng J et al (2022) exomePeak2: a peak calling and differential analysis tool for MeRIP-Seq with bias awareness. *NAR Genom Bioinform* 4(3):lqac046. doi:10.1093\u002Fnargab\u002Flqac046\n- McIntyre ABR, Gokhale NS, Cerchietti L, Jaffrey SR, Horner SM, Mason CE (2020) Limits in the detection of m6A changes using MeRIP\u002Fm6A-seq. *Sci Rep* 10(1):6590. doi:10.1038\u002Fs41598-020-63355-3\n- Liu C, Sun H, Yi Y et al (2023) Absolute quantification of single-base m6A methylation in the mammalian transcriptome using GLORI. *Nat Biotechnol* 41(3):355-366. doi:10.1038\u002Fs41587-022-01487-9\n- Hu L, Liu S, Peng Y et al (2022) m6A RNA modifications are measured at single-base resolution across the mammalian transcriptome. *Nat Biotechnol* 40(8):1210-1219. doi:10.1038\u002Fs41587-022-01243-z\n- Garcia-Campos MA, Edelheit S, Toth U et al (2019) Deciphering the m6A code via antibody-independent quantitative profiling. *Cell* 178(3):731-747.e16. doi:10.1016\u002Fj.cell.2019.06.013\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- 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- Liao Y, Smyth GK, Shi W (2014) featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. *Bioinformatics* 30(7):923-930. doi:10.1093\u002Fbioinformatics\u002Fbtt656\n\n## Related Skills\n\n- merip-preprocessing - Upstream IP\u002Finput BAM preparation; design-matrix metadata (antibody lot, batch) originates here\n- m6a-peak-calling - Peak calling step that produces input to differential analysis\n- m6anet-analysis - Orthogonal ONT-direct-RNA validation for stoichiometry claims at high-stakes loci\n- modification-visualization - Volcano \u002F MA \u002F per-peak boxplot rendering of differential results\n- differential-expression\u002Fdeseq2-basics - Canonical DE design philosophy; m6a-differential defers to this for general design-matrix patterns\n- differential-expression\u002Fde-results - Post-DE interpretation, ranking, gene-list extraction\n- differential-expression\u002Fedger-basics - edgeR fundamentals for the paired-symmetric sensitivity case\n- chip-seq\u002Fdifferential-binding - Closest sibling for IP-vs-input differential binding (general framework)\n- rna-quantification\u002Ffeaturecounts-counting - Peak count matrix construction\n- data-visualization\u002Fvolcano-and-ma-plots - Volcano + MA plot recipes\n- data-visualization\u002Fmultipanel-figures - Figure assembly\n- pathway-analysis\u002Fgo-enrichment - GO enrichment on differential-peak-bearing gene lists\n- workflows\u002Frnaseq-to-de - End-to-end pipeline orchestration patterns\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","epitranscriptomics\u002Fm6a-differential\u002FSKILL.md","exomePeak2","GPTomics\u002FbioSkills","main","2d1750717747097b602acf30bd09a7cbd4b4c6ea","awesome_skills\u002Fbiology\u002Fbio-epitranscriptomics-m6a-differential","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-epitranscriptomics-m6a-differential",{"tool":42,"command":43},"bunx","bunx skills add Mr-Nobody-Anonymous\u002FAll-skills --skill bio-epitranscriptomics-m6a-differential",{"tool":45,"command":46},"pnpm","pnpm dlx skills add Mr-Nobody-Anonymous\u002FAll-skills --skill bio-epitranscriptomics-m6a-differential","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fbiology\u002Fbio-epitranscriptomics-m6a-differential\u002FSKILL.md",true,[50,56],{"path":51,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":54,"sourceHtmlUrl":55},"examples\u002Fdifferential_m6a.R","OTHER","其他",3465,"https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fbiology\u002Fbio-epitranscriptomics-m6a-differential\u002Fexamples\u002Fdifferential_m6a.R",{"path":57,"resourceType":58,"resourceTypeDesc":59,"sizeBytes":60,"sourceHtmlUrl":61},"usage-guide.md","REFERENCE","参考文档",6884,"https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fbiology\u002Fbio-epitranscriptomics-m6a-differential\u002Fusage-guide.md",[63,71,78,85,92,99],{"skillNo":64,"name":65,"nameZh":6,"description":66,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":67,"categories":68,"keywords":69,"publishedAt":70},"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":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":70},"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":79,"name":80,"nameZh":6,"description":81,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":82,"categories":83,"keywords":84,"publishedAt":70},"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":86,"name":87,"nameZh":6,"description":88,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":89,"categories":90,"keywords":91,"publishedAt":70},"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":93,"name":94,"nameZh":6,"description":95,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":96,"categories":97,"keywords":98,"publishedAt":70},"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":100,"name":101,"nameZh":6,"description":102,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":103,"categories":104,"keywords":105,"publishedAt":70},"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",[],[]]