[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7EDSEAAQBKHQB0T3DA7NFJC":3,"skill-same:01M7EDSEAAQBKHQB0T3DA7NFJC":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},"01M7EDSEAAQBKHQB0T3DA7NFJC","bio-hi-c-analysis-loop-calling",null,"Detects focal chromatin loops (point interactions \u002F corner-dots) in balanced Hi-C and Micro-C contact maps and aggregates\u002Fvalidates a loop set. Covers de-novo calling with cooltools dots (HiCCUPS-style 4-background local enrichment with lambda-chunked FDR), chromosight (template-correlation), and Mustache (scale-space blob detection); aggregate peak analysis (APA) via cooltools pileup for confirmation; the depth\u002Fresolution prerequisite (de-novo needs ~5-10kb resolution = hundreds of millions to billions of valid pairs); consensus across callers and convergent-CTCF support as validation; and differential loops via union anchors plus chromosight quantify. Use when calling chromatin loops or dots from a cooler, deciding whether a map is deep enough to call de-novo vs running APA on known CTCF\u002Fcohesin anchors, building an aggregate peak pileup, comparing loops across conditions, or validating loop calls. For HiChIP\u002FPLAC-seq\u002FPCHi-C protein-anchored data use FitHiChIP\u002FMAPS, not dots.","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-hi-c-analysis-loop-calling\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: bio-hi-c-analysis-loop-calling\ndescription: Detects focal chromatin loops (point interactions \u002F corner-dots) in balanced Hi-C and Micro-C contact maps and aggregates\u002Fvalidates a loop set. Covers de-novo calling with cooltools dots (HiCCUPS-style 4-background local enrichment with lambda-chunked FDR), chromosight (template-correlation), and Mustache (scale-space blob detection); aggregate peak analysis (APA) via cooltools pileup for confirmation; the depth\u002Fresolution prerequisite (de-novo needs ~5-10kb resolution = hundreds of millions to billions of valid pairs); consensus across callers and convergent-CTCF support as validation; and differential loops via union anchors plus chromosight quantify. Use when calling chromatin loops or dots from a cooler, deciding whether a map is deep enough to call de-novo vs running APA on known CTCF\u002Fcohesin anchors, building an aggregate peak pileup, comparing loops across conditions, or validating loop calls. For HiChIP\u002FPLAC-seq\u002FPCHi-C protein-anchored data use FitHiChIP\u002FMAPS, not dots.\ntool_type: mixed\nprimary_tool: cooltools\nsource: \"https:\u002F\u002Fgithub.com\u002FGPTomics\u002FbioSkills\"\nsource_repository: \"GPTomics\u002FbioSkills\"\nsource_path: \"hi-c-analysis\u002Floop-calling\u002FSKILL.md\"\nlicense: \"MIT\"\nimported_at: \"2026-09-20\"\n---\n\n## Version Compatibility\n\nReference examples tested with: cooltools 0.7+, cooler 0.10+, bioframe 0.7+, chromosight 1.6+, mustache 1.3+\n\nBefore using code patterns, verify installed versions match. If versions differ:\n- Python: `pip show \u003Cpackage>` then `help(module.function)` to check signatures\n- CLI: `\u003Ctool> --version` then `\u003Ctool> --help` to confirm flags\n\nIf code throws ImportError, AttributeError, or TypeError, introspect the installed\npackage and adapt the example to match the actual API rather than retrying.\n\nThe `.cool` must be BALANCED before calling loops -- `dots`\u002F`pileup` read the `weight` column and raw counts are unsupported. An `.mcool` is multi-resolution; pass a single-resolution URI (`file.mcool::\u002Fresolutions\u002F10000`), not the bare `.mcool`. cooltools changed signatures around 0.5 -> 0.7 (view_df\u002Fexpected_value_col conventions); verify `help(cooltools.dots)` for the installed version. The `view_df` passed to `expected_cis` MUST be the same one passed to `dots`\u002F`pileup`.\n\n# Chromatin Loop Calling\n\n**\"Where are the focal loops (CTCF\u002Fcohesin corner-dots, E-P contacts) in my Hi-C map?\"** -> Test each off-diagonal pixel for focal enrichment against its local background (on a balanced, expected-normalized matrix), control FDR, then validate the set by aggregation and orthogonal support.\n- Python: `cooltools.dots(clr, expected=cooltools.expected_cis(clr, view_df=arms), view_df=arms)`\n- CLI: `chromosight detect --pattern loops --min-dist 20000 --max-dist 2000000 sample.cool::\u002Fresolutions\u002F5000 out`\n\n## The Single Most Important Modern Insight -- Loop Calling Is Depth-Limited, Not Algorithm-Limited; the First Question Is \"How Deep Is the Map?\"\n\nThe caller choice is second-order. The dominant variable in whether loops are found at all is sequencing depth \u002F map resolution. Rao 2014 needed ~4.9 BILLION contacts in GM12878 to reach 1kb bins and call ~10,000 loops; robust de-novo calling realistically wants 5-10kb resolution, which is hundreds of millions to billions of valid cis pairs. Below that, every caller returns near-nothing or noise, and tuning the FDR will not rescue it. So the workflow forks on depth before any tool is chosen:\n\n1. **Deep map (>=~500M-1B valid pairs, 5-10kb resolution):** de-novo calling is licensed. Run cooltools `dots` (or chromosight \u002F Mustache), then validate (see below).\n2. **Shallow map:** do NOT de-novo call. Run **APA \u002F pileup on a KNOWN anchor set** -- loops imported from a deep reference map, or anchor pairs built from CTCF\u002Fcohesin ChIP-seq peaks. This is the single most important practical reframe in the skill: shallow data can still *confirm and quantify* a hypothesized loop set even when it cannot *discover* one.\n\nTwo corollaries that follow directly:\n\n**De-novo calling DISCOVERS; APA CONFIRMS -- never conflate them.** APA aggregates many putative loops to surface mean signal no individual loop could pass FDR for. An enriched APA center pixel proves \"this SET of pairs is enriched on average\"; it does NOT prove any single pair is a loop and it cannot discover new loops. Presenting an APA pileup as evidence that \"these loops exist\" is the classic abuse. And the APA score is meaningless without a **corner control** -- center pixel divided by an off-diagonal corner block of the flank is the on-vs-off measurement; the bare center value alone says nothing.\n\n**Loops form between convergent CTCF motifs -- biology AND a validation filter.** Loops preferentially link two CTCF motifs in CONVERGENT orientation (Rao 2014 observation; de Wit 2015, Sanborn 2015 extrusion mechanism; proven by CTCF-site inversion experiments that kill or reroute the loop). A called corner-dot whose two anchors carry convergent CTCF motifs is high-confidence; one with no CTCF\u002Fanchor support on a shallow map is likely a false positive. Not all loops are CTCF loops (E-P and polycomb loops exist), so convergent-CTCF is a strong positive filter, not a universal requirement.\n\n## Loop-Caller Taxonomy\n\n| Tool | Philosophy | Mechanism | When |\n|------|-----------|-----------|------|\n| cooltools `dots` | local enrichment (CPU HiCCUPS) | pixel must beat 4 local backgrounds (donut\u002Fhorizontal\u002Fvertical\u002Flower-left); Poisson p; lambda-binned BH-FDR | cooler\u002F.mcool pipelines, the modern default; pure-CPU |\n| Juicer HiCCUPS | local enrichment (GPU original) | same 4-kernel model on `.hic`; CUDA-bound | `.hic`\u002FJuicer ecosystem with a GPU available |\n| chromosight `detect` | template correlation | Pearson correlation of a loop\u002Fborder\u002Fstripe kernel vs each window | want loops AND borders AND stripes from one engine; Micro-C-friendly |\n| Mustache | scale-space blobs | Difference-of-Gaussians across scales; multi-scale catches loops of different sizes | mixed loop sizes, kb-resolution Micro-C, recovers more E-P\u002FChIA-PET loops |\n| SIP | image processing | Gaussian blur + regional-max + watershed | `.hic` image-based alternative |\n| cooltools `pileup` (APA) | CONFIRMATION, not discovery | aggregate snippets centered on an anchor set; measure center vs corner | validate\u002Fquantify a loop SET; works on shallow maps |\n\nForcato 2017 (*Nat Methods* 14:679) is the canonical finding that loop callers show LOW pairwise overlap and poor replicate reproducibility -- far worse than TAD callers. Practical consequence: a loop called by only one tool is suspect. Trust comes from consensus across >=2 callers plus orthogonal support (convergent CTCF, ChIA-PET\u002FHiChIP), not from any single tool's list length.\n\n## Decision Tree by Scenario\n\n| Scenario | Recommended | Why |\n|----------|-------------|-----|\n| Shallow map (tens of M pairs) | APA\u002Fpileup on KNOWN anchors (CTCF\u002Fcohesin ChIP or reference loops) -- STOP de-novo | callers return noise below ~5-10kb resolution |\n| Deep cooler\u002F.mcool, CPU only | `cooltools.dots` (default) | pure-CPU HiCCUPS reimplementation on balanced cooler |\n| Deep `.hic` with a GPU | Juicer HiCCUPS | CUDA original built for billion-contact `.hic` scans |\n| Mixed loop sizes \u002F kb Micro-C | Mustache | scale-space natively spans loop sizes; sub-5kb-friendly |\n| Want stripes\u002Fborders too | chromosight (swap `--pattern`) | same template engine; stripes are a SEPARATE class, not loops |\n| Validate a call set | `cooltools.pileup` -> APA score vs corner control | aggregate enrichment + visual QC of the dot |\n| Confirm anchors are real loops | convergent-CTCF check -> chip-seq\u002Fpeak-annotation, atac-seq\u002Ffootprinting | extrusion loops carry convergent CTCF motifs |\n| Annotate loop anchors | -> chip-seq\u002Fpeak-annotation, atac-seq\u002Fenhancer-gene-linking | E-P \u002F TF context lives there |\n| Anchor-overlap enrichment p-value | -> genome-intervals\u002Foverlap-significance | turn an anchor-overlap count into a permutation test |\n| Two conditions, loop strength shift | union anchors -> chromosight `quantify` per condition -> test delta (or `diff_mustache`) | NO bin-level DESeq for loops; quantify a fixed coordinate set |\n| HiChIP \u002F PLAC-seq \u002F PCHi-C | FitHiChIP \u002F MAPS \u002F HiC-DC+ -> chip-seq\u002Fpeak-calling | protein-anchored, coverage-biased; HiCCUPS null is wrong |\n\n## De-Novo Loop Calling with cooltools dots\n\n**Goal:** Discover focal loops genome-wide on a deep, balanced map with honest FDR control.\n\n**Approach:** Build chromosome-arm regions, compute the distance-decay expected on those arms, then run `dots` -- which convolves four local-background kernels and runs Benjamini-Hochberg FDR independently within geometrically-spaced lambda-bins of locally-adjusted expected. The arms `view_df` must be identical for `expected_cis` and `dots`.\n\n```python\nimport cooler, cooltools, bioframe\n\nclr = cooler.Cooler('matrix.mcool::\u002Fresolutions\u002F10000')   # 10kb: a realistic de-novo floor; finer needs more depth\narms = bioframe.make_viewframe(clr.chromsizes)   # or cooltools.lib.read_viewframe_from_file('hg38_arms.bed', clr) for per-arm\nexpected = cooltools.expected_cis(clr, view_df=arms, nproc=4)   # distance-matched background; same view as dots\nloops = cooltools.dots(\n    clr, expected=expected, view_df=arms,\n    max_loci_separation=10_000_000,   # ignore pixels farther than 10Mb from the diagonal\n    n_lambda_bins=40, lambda_bin_fdr=0.1,   # FDR run independently per geometric lambda-bin (HiCCUPS default)\n    clustering_radius=20_000,   # merge called pixels within 20kb into one loop\n    nproc=4,\n)\n```\n\n**The four backgrounds, and why lower-left is the clever one.** A pixel must beat ALL four local-background kernels, not one. **Donut** = is it a focal enrichment at all. **Horizontal** and **vertical** = is it actually a STRIPE pixel masquerading as a dot (these kernels exist to NOT call architectural stripes as loops). **Lower-left** = is it just a TAD\u002Fcontact-domain CORNER -- a domain corner is enriched vs the donut but NOT vs its lower-left neighborhood, so requiring the pixel to also beat lower-left separates a genuine point loop from a generic domain corner. Skipping lower-left inflates calls with domain corners.\n\n**Lambda-chunking is why HiCCUPS FDR is honest.** Contact counts span orders of magnitude with genomic distance, so a single genome-wide BH-FDR would be dominated by the high-count near-diagonal regime and over-call. `dots` bins pixels by their locally-adjusted expected into geometrically-spaced lambda-bins (`n_lambda_bins=40`) and runs BH-FDR independently within each (`lambda_bin_fdr=0.1`), so low-count and high-count regimes are each thresholded correctly.\n\n## Template-Matching with chromosight\n\n```bash\n# detect: \u003Ccontact_map> \u003Cprefix> are positional and come LAST\nchromosight detect --pattern loops --threads 8 \\\n  --min-dist 20000 --max-dist 2000000 --pearson 0.4 \\\n  sample.cool::\u002Fresolutions\u002F5000 sample_loops\n# output: sample_loops.tsv -> chrom1,start1,end1,chrom2,start2,end2,bin1,bin2,score,pvalue,qvalue\n```\n\nThe score is a Pearson correlation (-1..1) between a loop kernel and each windowed submatrix. The same engine finds borders and stripes by swapping `--pattern` (loops, loops_small, borders, hairpins, centromeres, stripes_left, stripes_right) -- but stripes are a separate feature class, NOT loops. `--pearson` is the correlation cutoff; raise it for fewer, higher-confidence calls.\n\n## Scale-Space with Mustache\n\n```bash\nmustache -f sample.mcool -r 5000 -o loops.tsv -pt 0.1 -st 0.88 -norm weight -p 8\n# output: BIN1_CHR BIN1_START BIN1_END BIN2_CHR BIN2_START BIN2_END FDR DETECTION_SCALE\n```\n\n`-pt` is the FDR\u002Fp-value threshold (default 0.1), `-st` the sparsity filter (default 0.88), `-norm weight` for a balanced `.cool` (`KR` for `.hic`). Mustache spans loop sizes natively via Difference-of-Gaussians across scales, which is why it adapts to kb-resolution Micro-C better than fixed-kernel HiCCUPS.\n\n## Aggregate Peak Analysis (APA) -- Confirm, Don't Discover\n\n**Goal:** Quantify whether a loop SET is enriched on average and visually QC the call set (a clean aggregate dot = mostly real; a smeared\u002Fabsent center = contaminated).\n\n**Approach:** Compute expected, pile up observed\u002Fexpected snippets centered on each anchor pair, average across the stack, then report the APA score = center pixel divided by an off-diagonal corner-control block. Pass `expected_df` so snippets are O\u002FE and comparable across genomic separations.\n\n```python\nimport numpy as np\nimport cooltools\n\nexpected = cooltools.expected_cis(clr, view_df=arms, nproc=4)\nstack = cooltools.pileup(clr, loops, view_df=arms, expected_df=expected, flank=100_000, nproc=4)   # bedpe two-anchor features\napa = np.nanmean(stack, axis=0)   # pileup returns (n_snippets, D, D); average over axis 0 -> 2D aggregate\n\ncenter = apa.shape[0] \u002F\u002F 2\ncorner = 3   # 3x3 corner-control block (Rao 2014 lower-left convention)\napa_score = apa[center, center] \u002F np.nanmean(apa[-corner:, :corner])   # center vs lower-left corner; >1 = enriched\n```\n\n## Differential Loops -- Union Anchors, Not a Bin-Level Tool\n\n**Goal:** Find loops whose strength changes between conditions.\n\n**Approach:** There is NO DESeq-for-loops. Build a UNION anchor set across conditions, then score each loop's strength per condition at a FIXED coordinate set (`chromosight quantify`, which is purpose-built for this, or APA per condition), then test the strength delta.\n\n```bash\n# quantify scores a FIXED coordinate set; arg order: \u003Cbed2d> \u003Ccontact_map> \u003Cprefix>\nchromosight quantify --pattern loops union_anchors.bed2d condA.cool condA_q\nchromosight quantify --pattern loops union_anchors.bed2d condB.cool condB_q\n# compare the per-loop score columns; or diff_mustache.py -f1 A -f2 B -pt 0.05 -pt2 0.1 -r 5000 -o diff\n```\n\n`diffHic`, `multiHiCcompare`, and `dcHiC` operate on BINS or COMPARTMENTS, not focal loops -- do NOT use them as a loop-differential tool. Cross-reference hic-differential for the bin\u002Fcompartment regime.\n\n## Per-Method Failure Modes\n\n### De-novo calling on a shallow map\n**Trigger:** running `dots`\u002Fchromosight\u002FMustache on tens of millions of pairs or >=25kb bins. **Mechanism:** focal signal is below the noise floor without depth. **Symptom:** zero or a handful of scattered, irreproducible calls. **Fix:** STOP de-novo; run APA on a known anchor set (CTCF\u002Fcohesin ChIP or reference loops).\n\n### APA reported without a corner control\n**Trigger:** quoting the aggregate center-pixel value as the loop \"strength.\" **Mechanism:** without an off-diagonal corner the number has no on-vs-off baseline. **Symptom:** a \"high\" APA that reflects distance-decay, not looping. **Fix:** APA score = center \u002F corner-control block (Rao 2014 lower-left convention).\n\n### APA presented as proof loops exist\n**Trigger:** showing a pileup to claim \"these N loops are real.\" **Mechanism:** APA surfaces mean enrichment across a SET; it cannot validate any single loop or discover new ones. **Symptom:** confident per-loop claims backed only by an aggregate. **Fix:** treat APA as set-level confirmation; for per-loop confidence use consensus + convergent-CTCF.\n\n### Trusting a single caller's list\n**Trigger:** reporting \"Mustache found N loops\" with no cross-check. **Mechanism:** callers have low pairwise overlap (Forcato 2017). **Symptom:** a list that barely overlaps a second tool or replicate. **Fix:** intersect >=2 callers and require convergent-CTCF \u002F ChIA-PET \u002F HiChIP support.\n\n### Calling domain corners as loops\n**Trigger:** a caller without a lower-left background (or a custom kernel set). **Mechanism:** a TAD corner beats the donut but is not a point loop. **Symptom:** \"loops\" sitting exactly at TAD corners with no anchor support. **Fix:** use `dots` (it beats all four backgrounds); cross-check anchors.\n\n### Calling stripe pixels as dots\n**Trigger:** detecting on a map with strong architectural stripes. **Mechanism:** a stripe pixel is enriched vs the donut but lies on a horizontal\u002Fvertical band. **Symptom:** \"loops\" smeared along a row\u002Fcolumn. **Fix:** the horizontal\u002Fvertical kernels suppress these in `dots`; treat stripes as a separate class (chromosight stripes_*).\n\n### 10kb-tuned kernels on 1kb Micro-C\n**Trigger:** default HiCCUPS donut\u002Fpeak-width on sub-5kb Micro-C. **Mechanism:** kernels are sized for 5-10kb Hi-C. **Symptom:** blurred or missed fine E-P loops. **Fix:** shrink the kernels for sub-5kb, or use Mustache\u002Fchromosight which adapt more gracefully.\n\n### Raw (unbalanced) matrix into dots\n**Trigger:** `dots` on a cooler with no `weight` column. **Mechanism:** dots requires balancing weights + expected. **Symptom:** error or meaningless output. **Fix:** `cooler balance` first; confirm `clr.matrix(balance=True)` is not all-NaN.\n\n### HiCCUPS-style calling on HiChIP\u002FPLAC-seq\n**Trigger:** running `dots` on cohesin\u002FH3K27ac HiChIP or PLAC-seq. **Mechanism:** protein-anchored data is coverage-biased; the Hi-C null is wrong. **Symptom:** distorted FDR, wrong loop counts. **Fix:** use FitHiChIP\u002FMAPS\u002FHiC-DC+ against a protein-anchored background.\n\n## Quantitative Thresholds\n\n| Threshold | Source | Rationale |\n|-----------|--------|-----------|\n| De-novo loop resolution 5-10kb | Rao 2014 depth scaling | ~4.9B contacts reached 1kb \u002F ~10k loops; coarser bins blur anchors, shallow maps cannot resolve them |\n| `max_loci_separation` 2-10Mb | loop size distribution | most loops are \u003C2Mb; 10Mb is the cooltools default ceiling on diagonal distance |\n| `n_lambda_bins=40`, `lambda_bin_fdr=0.1` | cooltools\u002FHiCCUPS default | geometric lambda-binning + per-bin BH-FDR keeps FDR honest across the count dynamic range |\n| `clustering_radius=20_000` | cooltools default | merges adjacent called pixels into one loop call |\n| chromosight `--pearson` ~0.4 loops | chromosight default | template-correlation cutoff; raise for higher-confidence, fewer calls |\n| Mustache `-pt 0.1`, `-st 0.88` | Mustache defaults | p\u002FFDR threshold and sparsity filter |\n| Consensus across >=2 callers | Forcato 2017 low overlap | single-caller lists are unreliable; require intersection or orthogonal support |\n| APA score = center \u002F corner block | Rao 2014 | the corner is the on-vs-off control; the bare center is uninterpretable |\n\n## Common Errors\n\n| Error \u002F symptom | Cause | Solution |\n|-----------------|-------|----------|\n| `dots` returns nothing \u002F scattered junk | map too shallow or resolution too coarse | check depth; below ~5-10kb resolution run APA on known anchors instead |\n| `clr.matrix(balance=True)` all NaN | cooler not balanced | `cooler balance` \u002F `cooler.balance_cooler` before calling loops |\n| `expected`\u002F`dots` shape or view error | different `view_df` for expected vs dots | reuse the same `view_df` (arms) for `expected_cis` and `dots` |\n| Empty \u002F wrong-resolution result on `.mcool` | bare `.mcool` passed | use `file.mcool::\u002Fresolutions\u002F\u003Cbp>` URI |\n| Empty result, no error | chrom naming mismatch (`chr1` vs `1`) across cooler\u002Fanchors\u002Fpeaks | harmonize chromosome naming everywhere |\n| `AttributeError` on a cooltools function | pre-0.7 vs 0.7+ signature change | `help(cooltools.dots)`; adapt to the installed signature |\n| APA center looks high but loops are weak | no corner control \u002F `expected_df` omitted | pass `expected_df` and divide center by a corner block |\n\n## References\n\n- Rao SSP, Huntley MH, Durand NC, et al. 2014. A 3D map of the human genome at kilobase resolution reveals principles of chromatin looping. *Cell* 159(7):1665-1680.\n- Open2C, Abdennur N, Abraham S, Fudenberg G, et al. 2024. Cooltools: enabling high-resolution Hi-C analysis in Python. *PLoS Comput Biol* 20(5):e1012067.\n- Matthey-Doret C, Baudry L, Breuer A, et al. 2020. Computer vision for pattern detection in chromosome contact maps (chromosight). *Nat Commun* 11:5795.\n- Roayaei Ardakany A, Gezer HT, Lonardi S, Ay F. 2020. Mustache: multi-scale detection of chromatin loops from Hi-C and Micro-C maps using scale-space representation. *Genome Biol* 21:256.\n- Rowley MJ, Poulet A, Nichols MH, et al. 2020. Analysis of Hi-C data using SIP effectively identifies loops in organisms from C. elegans to mammals. *Genome Res* 30(3):447-458.\n- Forcato M, Nicoletti C, Pal K, et al. 2017. Comparison of computational methods for Hi-C data analysis. *Nat Methods* 14:679-685.\n- de Wit E, Vos ESM, Holwerda SJB, et al. 2015. CTCF binding polarity determines chromatin looping. *Mol Cell* 60(4):676-684.\n- Sanborn AL, Rao SSP, Huang SC, et al. 2015. Chromatin extrusion explains key features of loop and domain formation. *PNAS* 112(47):E6456-E6465.\n- Rao SSP, Huang SC, Glenn St Hilaire B, et al. 2017. Cohesin loss eliminates all loop domains. *Cell* 171(2):305-320.\n- Haarhuis JHI, van der Weide RH, Blomen VA, et al. 2017. The cohesin release factor WAPL restricts chromatin loop extension. *Cell* 169(4):693-707.\n- Schwarzer W, Abdennur N, Goloborodko A, et al. 2017. Two independent modes of chromatin organization revealed by cohesin removal. *Nature* 551:51-56.\n- Krietenstein N, Abraham S, Venev SV, et al. 2020. Ultrastructural details of mammalian chromosome architecture (Micro-C). *Mol Cell* 78(3):554-565.\n- Hsieh THS, Cattoglio C, Slobodyanyuk E, et al. 2020. Resolving the 3D landscape of transcription-linked mammalian chromatin folding (Micro-C). *Mol Cell* 78(3):539-553.\n- Bhattacharyya S, Chandra V, Vijayanand P, Ay F. 2019. Identification of significant chromatin contacts from HiChIP data by FitHiChIP. *Nat Commun* 10:4221.\n\n## Related Skills\n\n- hic-data-io - Load and access the cooler files this skill calls loops on\n- matrix-operations - Balancing and expected\u002FO\u002FE that dots and pileup depend on\n- hic-visualization - Render called loops and APA pileups on the heatmap\n- hic-differential - Bin\u002Fcompartment-level differential (the regime loops are NOT in)\n- tad-detection - TAD corners vs point loops; the lower-left background separates them\n- chip-seq\u002Fpeak-calling - CTCF\u002Fcohesin peaks to anchor and validate loops; HiChIP peak context\n- chip-seq\u002Fpeak-annotation - Annotate loop anchors with TF\u002FCTCF peaks\n- atac-seq\u002Fenhancer-gene-linking - E-P contacts complementing loop calls\n- atac-seq\u002Ffootprinting - TF footprints at loop anchors\n- genome-intervals\u002Foverlap-significance - Permutation test for anchor\u002Ffeature enrichment\n","## Version Compatibility\n\nReference examples tested with: cooltools 0.7+, cooler 0.10+, bioframe 0.7+, chromosight 1.6+, mustache 1.3+\n\nBefore using code patterns, verify installed versions match. If versions differ:\n- Python: `pip show \u003Cpackage>` then `help(module.function)` to check signatures\n- CLI: `\u003Ctool> --version` then `\u003Ctool> --help` to confirm flags\n\nIf code throws ImportError, AttributeError, or TypeError, introspect the installed\npackage and adapt the example to match the actual API rather than retrying.\n\nThe `.cool` must be BALANCED before calling loops -- `dots`\u002F`pileup` read the `weight` column and raw counts are unsupported. An `.mcool` is multi-resolution; pass a single-resolution URI (`file.mcool::\u002Fresolutions\u002F10000`), not the bare `.mcool`. cooltools changed signatures around 0.5 -> 0.7 (view_df\u002Fexpected_value_col conventions); verify `help(cooltools.dots)` for the installed version. The `view_df` passed to `expected_cis` MUST be the same one passed to `dots`\u002F`pileup`.\n\n# Chromatin Loop Calling\n\n**\"Where are the focal loops (CTCF\u002Fcohesin corner-dots, E-P contacts) in my Hi-C map?\"** -> Test each off-diagonal pixel for focal enrichment against its local background (on a balanced, expected-normalized matrix), control FDR, then validate the set by aggregation and orthogonal support.\n- Python: `cooltools.dots(clr, expected=cooltools.expected_cis(clr, view_df=arms), view_df=arms)`\n- CLI: `chromosight detect --pattern loops --min-dist 20000 --max-dist 2000000 sample.cool::\u002Fresolutions\u002F5000 out`\n\n## The Single Most Important Modern Insight -- Loop Calling Is Depth-Limited, Not Algorithm-Limited; the First Question Is \"How Deep Is the Map?\"\n\nThe caller choice is second-order. The dominant variable in whether loops are found at all is sequencing depth \u002F map resolution. Rao 2014 needed ~4.9 BILLION contacts in GM12878 to reach 1kb bins and call ~10,000 loops; robust de-novo calling realistically wants 5-10kb resolution, which is hundreds of millions to billions of valid cis pairs. Below that, every caller returns near-nothing or noise, and tuning the FDR will not rescue it. So the workflow forks on depth before any tool is chosen:\n\n1. **Deep map (>=~500M-1B valid pairs, 5-10kb resolution):** de-novo calling is licensed. Run cooltools `dots` (or chromosight \u002F Mustache), then validate (see below).\n2. **Shallow map:** do NOT de-novo call. Run **APA \u002F pileup on a KNOWN anchor set** -- loops imported from a deep reference map, or anchor pairs built from CTCF\u002Fcohesin ChIP-seq peaks. This is the single most important practical reframe in the skill: shallow data can still *confirm and quantify* a hypothesized loop set even when it cannot *discover* one.\n\nTwo corollaries that follow directly:\n\n**De-novo calling DISCOVERS; APA CONFIRMS -- never conflate them.** APA aggregates many putative loops to surface mean signal no individual loop could pass FDR for. An enriched APA center pixel proves \"this SET of pairs is enriched on average\"; it does NOT prove any single pair is a loop and it cannot discover new loops. Presenting an APA pileup as evidence that \"these loops exist\" is the classic abuse. And the APA score is meaningless without a **corner control** -- center pixel divided by an off-diagonal corner block of the flank is the on-vs-off measurement; the bare center value alone says nothing.\n\n**Loops form between convergent CTCF motifs -- biology AND a validation filter.** Loops preferentially link two CTCF motifs in CONVERGENT orientation (Rao 2014 observation; de Wit 2015, Sanborn 2015 extrusion mechanism; proven by CTCF-site inversion experiments that kill or reroute the loop). A called corner-dot whose two anchors carry convergent CTCF motifs is high-confidence; one with no CTCF\u002Fanchor support on a shallow map is likely a false positive. Not all loops are CTCF loops (E-P and polycomb loops exist), so convergent-CTCF is a strong positive filter, not a universal requirement.\n\n## Loop-Caller Taxonomy\n\n| Tool | Philosophy | Mechanism | When |\n|------|-----------|-----------|------|\n| cooltools `dots` | local enrichment (CPU HiCCUPS) | pixel must beat 4 local backgrounds (donut\u002Fhorizontal\u002Fvertical\u002Flower-left); Poisson p; lambda-binned BH-FDR | cooler\u002F.mcool pipelines, the modern default; pure-CPU |\n| Juicer HiCCUPS | local enrichment (GPU original) | same 4-kernel model on `.hic`; CUDA-bound | `.hic`\u002FJuicer ecosystem with a GPU available |\n| chromosight `detect` | template correlation | Pearson correlation of a loop\u002Fborder\u002Fstripe kernel vs each window | want loops AND borders AND stripes from one engine; Micro-C-friendly |\n| Mustache | scale-space blobs | Difference-of-Gaussians across scales; multi-scale catches loops of different sizes | mixed loop sizes, kb-resolution Micro-C, recovers more E-P\u002FChIA-PET loops |\n| SIP | image processing | Gaussian blur + regional-max + watershed | `.hic` image-based alternative |\n| cooltools `pileup` (APA) | CONFIRMATION, not discovery | aggregate snippets centered on an anchor set; measure center vs corner | validate\u002Fquantify a loop SET; works on shallow maps |\n\nForcato 2017 (*Nat Methods* 14:679) is the canonical finding that loop callers show LOW pairwise overlap and poor replicate reproducibility -- far worse than TAD callers. Practical consequence: a loop called by only one tool is suspect. Trust comes from consensus across >=2 callers plus orthogonal support (convergent CTCF, ChIA-PET\u002FHiChIP), not from any single tool's list length.\n\n## Decision Tree by Scenario\n\n| Scenario | Recommended | Why |\n|----------|-------------|-----|\n| Shallow map (tens of M pairs) | APA\u002Fpileup on KNOWN anchors (CTCF\u002Fcohesin ChIP or reference loops) -- STOP de-novo | callers return noise below ~5-10kb resolution |\n| Deep cooler\u002F.mcool, CPU only | `cooltools.dots` (default) | pure-CPU HiCCUPS reimplementation on balanced cooler |\n| Deep `.hic` with a GPU | Juicer HiCCUPS | CUDA original built for billion-contact `.hic` scans |\n| Mixed loop sizes \u002F kb Micro-C | Mustache | scale-space natively spans loop sizes; sub-5kb-friendly |\n| Want stripes\u002Fborders too | chromosight (swap `--pattern`) | same template engine; stripes are a SEPARATE class, not loops |\n| Validate a call set | `cooltools.pileup` -> APA score vs corner control | aggregate enrichment + visual QC of the dot |\n| Confirm anchors are real loops | convergent-CTCF check -> chip-seq\u002Fpeak-annotation, atac-seq\u002Ffootprinting | extrusion loops carry convergent CTCF motifs |\n| Annotate loop anchors | -> chip-seq\u002Fpeak-annotation, atac-seq\u002Fenhancer-gene-linking | E-P \u002F TF context lives there |\n| Anchor-overlap enrichment p-value | -> genome-intervals\u002Foverlap-significance | turn an anchor-overlap count into a permutation test |\n| Two conditions, loop strength shift | union anchors -> chromosight `quantify` per condition -> test delta (or `diff_mustache`) | NO bin-level DESeq for loops; quantify a fixed coordinate set |\n| HiChIP \u002F PLAC-seq \u002F PCHi-C | FitHiChIP \u002F MAPS \u002F HiC-DC+ -> chip-seq\u002Fpeak-calling | protein-anchored, coverage-biased; HiCCUPS null is wrong |\n\n## De-Novo Loop Calling with cooltools dots\n\n**Goal:** Discover focal loops genome-wide on a deep, balanced map with honest FDR control.\n\n**Approach:** Build chromosome-arm regions, compute the distance-decay expected on those arms, then run `dots` -- which convolves four local-background kernels and runs Benjamini-Hochberg FDR independently within geometrically-spaced lambda-bins of locally-adjusted expected. The arms `view_df` must be identical for `expected_cis` and `dots`.\n\n```python\nimport cooler, cooltools, bioframe\n\nclr = cooler.Cooler('matrix.mcool::\u002Fresolutions\u002F10000')   # 10kb: a realistic de-novo floor; finer needs more depth\narms = bioframe.make_viewframe(clr.chromsizes)   # or cooltools.lib.read_viewframe_from_file('hg38_arms.bed', clr) for per-arm\nexpected = cooltools.expected_cis(clr, view_df=arms, nproc=4)   # distance-matched background; same view as dots\nloops = cooltools.dots(\n    clr, expected=expected, view_df=arms,\n    max_loci_separation=10_000_000,   # ignore pixels farther than 10Mb from the diagonal\n    n_lambda_bins=40, lambda_bin_fdr=0.1,   # FDR run independently per geometric lambda-bin (HiCCUPS default)\n    clustering_radius=20_000,   # merge called pixels within 20kb into one loop\n    nproc=4,\n)\n```\n\n**The four backgrounds, and why lower-left is the clever one.** A pixel must beat ALL four local-background kernels, not one. **Donut** = is it a focal enrichment at all. **Horizontal** and **vertical** = is it actually a STRIPE pixel masquerading as a dot (these kernels exist to NOT call architectural stripes as loops). **Lower-left** = is it just a TAD\u002Fcontact-domain CORNER -- a domain corner is enriched vs the donut but NOT vs its lower-left neighborhood, so requiring the pixel to also beat lower-left separates a genuine point loop from a generic domain corner. Skipping lower-left inflates calls with domain corners.\n\n**Lambda-chunking is why HiCCUPS FDR is honest.** Contact counts span orders of magnitude with genomic distance, so a single genome-wide BH-FDR would be dominated by the high-count near-diagonal regime and over-call. `dots` bins pixels by their locally-adjusted expected into geometrically-spaced lambda-bins (`n_lambda_bins=40`) and runs BH-FDR independently within each (`lambda_bin_fdr=0.1`), so low-count and high-count regimes are each thresholded correctly.\n\n## Template-Matching with chromosight\n\n```bash\n# detect: \u003Ccontact_map> \u003Cprefix> are positional and come LAST\nchromosight detect --pattern loops --threads 8 \\\n  --min-dist 20000 --max-dist 2000000 --pearson 0.4 \\\n  sample.cool::\u002Fresolutions\u002F5000 sample_loops\n# output: sample_loops.tsv -> chrom1,start1,end1,chrom2,start2,end2,bin1,bin2,score,pvalue,qvalue\n```\n\nThe score is a Pearson correlation (-1..1) between a loop kernel and each windowed submatrix. The same engine finds borders and stripes by swapping `--pattern` (loops, loops_small, borders, hairpins, centromeres, stripes_left, stripes_right) -- but stripes are a separate feature class, NOT loops. `--pearson` is the correlation cutoff; raise it for fewer, higher-confidence calls.\n\n## Scale-Space with Mustache\n\n```bash\nmustache -f sample.mcool -r 5000 -o loops.tsv -pt 0.1 -st 0.88 -norm weight -p 8\n# output: BIN1_CHR BIN1_START BIN1_END BIN2_CHR BIN2_START BIN2_END FDR DETECTION_SCALE\n```\n\n`-pt` is the FDR\u002Fp-value threshold (default 0.1), `-st` the sparsity filter (default 0.88), `-norm weight` for a balanced `.cool` (`KR` for `.hic`). Mustache spans loop sizes natively via Difference-of-Gaussians across scales, which is why it adapts to kb-resolution Micro-C better than fixed-kernel HiCCUPS.\n\n## Aggregate Peak Analysis (APA) -- Confirm, Don't Discover\n\n**Goal:** Quantify whether a loop SET is enriched on average and visually QC the call set (a clean aggregate dot = mostly real; a smeared\u002Fabsent center = contaminated).\n\n**Approach:** Compute expected, pile up observed\u002Fexpected snippets centered on each anchor pair, average across the stack, then report the APA score = center pixel divided by an off-diagonal corner-control block. Pass `expected_df` so snippets are O\u002FE and comparable across genomic separations.\n\n```python\nimport numpy as np\nimport cooltools\n\nexpected = cooltools.expected_cis(clr, view_df=arms, nproc=4)\nstack = cooltools.pileup(clr, loops, view_df=arms, expected_df=expected, flank=100_000, nproc=4)   # bedpe two-anchor features\napa = np.nanmean(stack, axis=0)   # pileup returns (n_snippets, D, D); average over axis 0 -> 2D aggregate\n\ncenter = apa.shape[0] \u002F\u002F 2\ncorner = 3   # 3x3 corner-control block (Rao 2014 lower-left convention)\napa_score = apa[center, center] \u002F np.nanmean(apa[-corner:, :corner])   # center vs lower-left corner; >1 = enriched\n```\n\n## Differential Loops -- Union Anchors, Not a Bin-Level Tool\n\n**Goal:** Find loops whose strength changes between conditions.\n\n**Approach:** There is NO DESeq-for-loops. Build a UNION anchor set across conditions, then score each loop's strength per condition at a FIXED coordinate set (`chromosight quantify`, which is purpose-built for this, or APA per condition), then test the strength delta.\n\n```bash\n# quantify scores a FIXED coordinate set; arg order: \u003Cbed2d> \u003Ccontact_map> \u003Cprefix>\nchromosight quantify --pattern loops union_anchors.bed2d condA.cool condA_q\nchromosight quantify --pattern loops union_anchors.bed2d condB.cool condB_q\n# compare the per-loop score columns; or diff_mustache.py -f1 A -f2 B -pt 0.05 -pt2 0.1 -r 5000 -o diff\n```\n\n`diffHic`, `multiHiCcompare`, and `dcHiC` operate on BINS or COMPARTMENTS, not focal loops -- do NOT use them as a loop-differential tool. Cross-reference hic-differential for the bin\u002Fcompartment regime.\n\n## Per-Method Failure Modes\n\n### De-novo calling on a shallow map\n**Trigger:** running `dots`\u002Fchromosight\u002FMustache on tens of millions of pairs or >=25kb bins. **Mechanism:** focal signal is below the noise floor without depth. **Symptom:** zero or a handful of scattered, irreproducible calls. **Fix:** STOP de-novo; run APA on a known anchor set (CTCF\u002Fcohesin ChIP or reference loops).\n\n### APA reported without a corner control\n**Trigger:** quoting the aggregate center-pixel value as the loop \"strength.\" **Mechanism:** without an off-diagonal corner the number has no on-vs-off baseline. **Symptom:** a \"high\" APA that reflects distance-decay, not looping. **Fix:** APA score = center \u002F corner-control block (Rao 2014 lower-left convention).\n\n### APA presented as proof loops exist\n**Trigger:** showing a pileup to claim \"these N loops are real.\" **Mechanism:** APA surfaces mean enrichment across a SET; it cannot validate any single loop or discover new ones. **Symptom:** confident per-loop claims backed only by an aggregate. **Fix:** treat APA as set-level confirmation; for per-loop confidence use consensus + convergent-CTCF.\n\n### Trusting a single caller's list\n**Trigger:** reporting \"Mustache found N loops\" with no cross-check. **Mechanism:** callers have low pairwise overlap (Forcato 2017). **Symptom:** a list that barely overlaps a second tool or replicate. **Fix:** intersect >=2 callers and require convergent-CTCF \u002F ChIA-PET \u002F HiChIP support.\n\n### Calling domain corners as loops\n**Trigger:** a caller without a lower-left background (or a custom kernel set). **Mechanism:** a TAD corner beats the donut but is not a point loop. **Symptom:** \"loops\" sitting exactly at TAD corners with no anchor support. **Fix:** use `dots` (it beats all four backgrounds); cross-check anchors.\n\n### Calling stripe pixels as dots\n**Trigger:** detecting on a map with strong architectural stripes. **Mechanism:** a stripe pixel is enriched vs the donut but lies on a horizontal\u002Fvertical band. **Symptom:** \"loops\" smeared along a row\u002Fcolumn. **Fix:** the horizontal\u002Fvertical kernels suppress these in `dots`; treat stripes as a separate class (chromosight stripes_*).\n\n### 10kb-tuned kernels on 1kb Micro-C\n**Trigger:** default HiCCUPS donut\u002Fpeak-width on sub-5kb Micro-C. **Mechanism:** kernels are sized for 5-10kb Hi-C. **Symptom:** blurred or missed fine E-P loops. **Fix:** shrink the kernels for sub-5kb, or use Mustache\u002Fchromosight which adapt more gracefully.\n\n### Raw (unbalanced) matrix into dots\n**Trigger:** `dots` on a cooler with no `weight` column. **Mechanism:** dots requires balancing weights + expected. **Symptom:** error or meaningless output. **Fix:** `cooler balance` first; confirm `clr.matrix(balance=True)` is not all-NaN.\n\n### HiCCUPS-style calling on HiChIP\u002FPLAC-seq\n**Trigger:** running `dots` on cohesin\u002FH3K27ac HiChIP or PLAC-seq. **Mechanism:** protein-anchored data is coverage-biased; the Hi-C null is wrong. **Symptom:** distorted FDR, wrong loop counts. **Fix:** use FitHiChIP\u002FMAPS\u002FHiC-DC+ against a protein-anchored background.\n\n## Quantitative Thresholds\n\n| Threshold | Source | Rationale |\n|-----------|--------|-----------|\n| De-novo loop resolution 5-10kb | Rao 2014 depth scaling | ~4.9B contacts reached 1kb \u002F ~10k loops; coarser bins blur anchors, shallow maps cannot resolve them |\n| `max_loci_separation` 2-10Mb | loop size distribution | most loops are \u003C2Mb; 10Mb is the cooltools default ceiling on diagonal distance |\n| `n_lambda_bins=40`, `lambda_bin_fdr=0.1` | cooltools\u002FHiCCUPS default | geometric lambda-binning + per-bin BH-FDR keeps FDR honest across the count dynamic range |\n| `clustering_radius=20_000` | cooltools default | merges adjacent called pixels into one loop call |\n| chromosight `--pearson` ~0.4 loops | chromosight default | template-correlation cutoff; raise for higher-confidence, fewer calls |\n| Mustache `-pt 0.1`, `-st 0.88` | Mustache defaults | p\u002FFDR threshold and sparsity filter |\n| Consensus across >=2 callers | Forcato 2017 low overlap | single-caller lists are unreliable; require intersection or orthogonal support |\n| APA score = center \u002F corner block | Rao 2014 | the corner is the on-vs-off control; the bare center is uninterpretable |\n\n## Common Errors\n\n| Error \u002F symptom | Cause | Solution |\n|-----------------|-------|----------|\n| `dots` returns nothing \u002F scattered junk | map too shallow or resolution too coarse | check depth; below ~5-10kb resolution run APA on known anchors instead |\n| `clr.matrix(balance=True)` all NaN | cooler not balanced | `cooler balance` \u002F `cooler.balance_cooler` before calling loops |\n| `expected`\u002F`dots` shape or view error | different `view_df` for expected vs dots | reuse the same `view_df` (arms) for `expected_cis` and `dots` |\n| Empty \u002F wrong-resolution result on `.mcool` | bare `.mcool` passed | use `file.mcool::\u002Fresolutions\u002F\u003Cbp>` URI |\n| Empty result, no error | chrom naming mismatch (`chr1` vs `1`) across cooler\u002Fanchors\u002Fpeaks | harmonize chromosome naming everywhere |\n| `AttributeError` on a cooltools function | pre-0.7 vs 0.7+ signature change | `help(cooltools.dots)`; adapt to the installed signature |\n| APA center looks high but loops are weak | no corner control \u002F `expected_df` omitted | pass `expected_df` and divide center by a corner block |\n\n## References\n\n- Rao SSP, Huntley MH, Durand NC, et al. 2014. A 3D map of the human genome at kilobase resolution reveals principles of chromatin looping. *Cell* 159(7):1665-1680.\n- Open2C, Abdennur N, Abraham S, Fudenberg G, et al. 2024. Cooltools: enabling high-resolution Hi-C analysis in Python. *PLoS Comput Biol* 20(5):e1012067.\n- Matthey-Doret C, Baudry L, Breuer A, et al. 2020. Computer vision for pattern detection in chromosome contact maps (chromosight). *Nat Commun* 11:5795.\n- Roayaei Ardakany A, Gezer HT, Lonardi S, Ay F. 2020. Mustache: multi-scale detection of chromatin loops from Hi-C and Micro-C maps using scale-space representation. *Genome Biol* 21:256.\n- Rowley MJ, Poulet A, Nichols MH, et al. 2020. Analysis of Hi-C data using SIP effectively identifies loops in organisms from C. elegans to mammals. *Genome Res* 30(3):447-458.\n- Forcato M, Nicoletti C, Pal K, et al. 2017. Comparison of computational methods for Hi-C data analysis. *Nat Methods* 14:679-685.\n- de Wit E, Vos ESM, Holwerda SJB, et al. 2015. CTCF binding polarity determines chromatin looping. *Mol Cell* 60(4):676-684.\n- Sanborn AL, Rao SSP, Huang SC, et al. 2015. Chromatin extrusion explains key features of loop and domain formation. *PNAS* 112(47):E6456-E6465.\n- Rao SSP, Huang SC, Glenn St Hilaire B, et al. 2017. Cohesin loss eliminates all loop domains. *Cell* 171(2):305-320.\n- Haarhuis JHI, van der Weide RH, Blomen VA, et al. 2017. The cohesin release factor WAPL restricts chromatin loop extension. *Cell* 169(4):693-707.\n- Schwarzer W, Abdennur N, Goloborodko A, et al. 2017. Two independent modes of chromatin organization revealed by cohesin removal. *Nature* 551:51-56.\n- Krietenstein N, Abraham S, Venev SV, et al. 2020. Ultrastructural details of mammalian chromosome architecture (Micro-C). *Mol Cell* 78(3):554-565.\n- Hsieh THS, Cattoglio C, Slobodyanyuk E, et al. 2020. Resolving the 3D landscape of transcription-linked mammalian chromatin folding (Micro-C). *Mol Cell* 78(3):539-553.\n- Bhattacharyya S, Chandra V, Vijayanand P, Ay F. 2019. Identification of significant chromatin contacts from HiChIP data by FitHiChIP. *Nat Commun* 10:4221.\n\n## Related Skills\n\n- hic-data-io - Load and access the cooler files this skill calls loops on\n- matrix-operations - Balancing and expected\u002FO\u002FE that dots and pileup depend on\n- hic-visualization - Render called loops and APA pileups on the heatmap\n- hic-differential - Bin\u002Fcompartment-level differential (the regime loops are NOT in)\n- tad-detection - TAD corners vs point loops; the lower-left background separates them\n- chip-seq\u002Fpeak-calling - CTCF\u002Fcohesin peaks to anchor and validate loops; HiChIP peak context\n- chip-seq\u002Fpeak-annotation - Annotate loop anchors with TF\u002FCTCF peaks\n- atac-seq\u002Fenhancer-gene-linking - E-P contacts complementing loop calls\n- atac-seq\u002Ffootprinting - TF footprints at loop anchors\n- genome-intervals\u002Foverlap-significance - Permutation test for anchor\u002Ffeature enrichment\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","mixed","2026-09-20","hi-c-analysis\u002Floop-calling\u002FSKILL.md","cooltools","GPTomics\u002FbioSkills","main","2d1750717747097b602acf30bd09a7cbd4b4c6ea","awesome_skills\u002Fbiology\u002Fbio-hi-c-analysis-loop-calling","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-hi-c-analysis-loop-calling",{"tool":42,"command":43},"bunx","bunx skills add Mr-Nobody-Anonymous\u002FAll-skills --skill bio-hi-c-analysis-loop-calling",{"tool":45,"command":46},"pnpm","pnpm dlx skills add Mr-Nobody-Anonymous\u002FAll-skills --skill bio-hi-c-analysis-loop-calling","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fbiology\u002Fbio-hi-c-analysis-loop-calling\u002FSKILL.md",true,[50,56],{"path":51,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":54,"sourceHtmlUrl":55},"examples\u002Fcall_loops.py","SCRIPT","脚本",2594,"https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fbiology\u002Fbio-hi-c-analysis-loop-calling\u002Fexamples\u002Fcall_loops.py",{"path":57,"resourceType":58,"resourceTypeDesc":59,"sizeBytes":60,"sourceHtmlUrl":61},"usage-guide.md","REFERENCE","参考文档",5584,"https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fbiology\u002Fbio-hi-c-analysis-loop-calling\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",[],[]]