Fine Mapping

🎯 SuSiE Fine-Mapper

You are SuSiE Fine-Mapper, a specialised ClawBio agent for statistical fine-mapping of GWAS loci. Your role is to identify credible sets of likely causal variants and compute per-variant posterior inclusion probabilities (PIPs) from GWAS summary statistics.

Why This Exists

GWAS identifies associated loci, not causal variants. A single GWAS signal can contain dozens of correlated SNPs in high LD — fine-mapping colocalises the signal onto the minimal credible set of likely causal variants.

• Without it: Researchers must manually triage 10–200 correlated SNPs per locus with no principled prioritisation
• With it: A ranked credible set with PIPs and 95% credible set boundaries in seconds
• Why ClawBio: Runs locally without uploading individual-level data; implements ABF natively and wraps SuSiE (via polyfun) when available — no R dependency required

Core Capabilities

1. Approximate Bayes Factors (ABF): Single-causal-variant fine-mapping from z-scores alone; no LD matrix required
2. SuSiE (Sum of Single Effects): Multi-signal fine-mapping with LD using the iterative Bayesian stepwise selection algorithm; pure-Python implementation, no R dependency
3. SuSiE-inf: SuSiE extended with an infinitesimal polygenic background component (τ²); produces tighter credible sets at well-powered loci by absorbing diffuse background signal; recommended when N > 50k or locus shows residual polygenic inflation
4. Swappable benchmark: tests/benchmark/finemapping_benchmark.py evaluates ABF, SuSiE, and SuSiE-inf head-to-head on synthetic loci with known causal variants; composite score (recall, precision, PIP concentration, rank)
5. Credible sets: 95% and 99% credible sets computed from PIPs; reports size, coverage, and lead variant
6. Visualisation: Locus PIP plot (colour-coded by LD r²), regional association plot overlaid with PIPs (optionally with a gene track fetched from Ensembl), credible set summary table
7. LD computation: Accepts a pre-computed LD matrix (.npy or .tsv)

Input Formats

| Format | Extension | Required Fields | Example | |--------|-----------|-----------------|---------| | GWAS summary stats | .tsv / .csv / .txt | rsid, chr, pos, beta, se or z | locus_sumstats.tsv | | Pre-computed LD matrix | .npy / .tsv | Square correlation matrix, row/col = variant order | ld_matrix.npy | | Demo (built-in) | — | — | --demo |

Optional columns in sumstats: p, maf, n, a1, a2

Workflow

When the user asks for fine-mapping:

1. Parse: Load sumstats TSV; detect z-score vs beta+se input; filter to locus window if --chr/--start/--end provided
2. LD: If --ld matrix supplied, load and validate dimensions match variants; if neither, run ABF (no LD needed)
3. Fine-map: Run ABF for single-signal or SuSiE for multi-signal; compute PIPs and credible sets
4. Visualise: Generate locus PIP plot; colour variants by LD r² to lead variant
5. Report: Write report.md with credible set tables, PIPs, methodology note, and reproducibility bundle

CLI Reference

# ABF single-signal fine-mapping (no LD needed)
python skills/fine-mapping/fine_mapping.py \
  --sumstats locus.tsv --output /tmp/finemapping

# SuSiE multi-signal with pre-computed LD matrix
python skills/fine-mapping/fine_mapping.py \
  --sumstats locus.tsv --ld ld_matrix.npy --output /tmp/finemapping

# Filter to a specific locus window
python skills/fine-mapping/fine_mapping.py \
  --sumstats gwas_full.tsv --chr 1 --start 109000000 --end 110000000 \
  --ld ld_matrix.npy --output /tmp/finemapping

# Set maximum number of causal signals (SuSiE L parameter)
python skills/fine-mapping/fine_mapping.py \
  --sumstats locus.tsv --ld ld_matrix.npy --max-signals 5 --output /tmp/finemapping

# Add a gene track below the regional association plot (requires internet)
python skills/fine-mapping/fine_mapping.py \
  --sumstats locus.tsv --ld ld_matrix.npy --gene-track --output /tmp/finemapping

# Demo mode (synthetic 200-variant locus, two causal signals)
python skills/fine-mapping/fine_mapping.py --demo --output /tmp/finemapping_demo

Demo

python skills/fine-mapping/fine_mapping.py --demo --output /tmp/finemapping_demo

Expected output: a report covering a synthetic 200-variant locus with two injected causal signals, SuSiE credible sets of ~3–8 variants each, per-variant PIP plot, and reproducibility bundle.

Algorithm / Methodology

Approximate Bayes Factors (ABF)

Used when no LD matrix is available (assumes variants are independent).

For each variant i with z-score z_i and prior variance W:

V_i  = 1 / n_eff    (if se available: V_i = se_i^2)
ABF_i = sqrt(V_i / (V_i + W)) * exp(z_i^2 * W / (2 * (V_i + W)))
PIP_i = ABF_i / sum(ABF_j)

Default prior: W = 0.04 (σ = 0.2 on log-OR scale; Wakefield 2009)

SuSiE (Sum of Single Effects, Wang et al. 2020)

When an LD matrix R is provided:

1. Initialise L single-effect vectors α_l (L = number of expected causal signals, default 10)
2. Iterative Bayesian Stepwise Selection (IBSS):

• For each effect l, compute residual z-scores removing all other effects
• Update α_l via single-effect regression posterior: α_l ∝ ABF(z_residual | R)
• Update posterior variance μ_l² and σ_l²

1. Converge when ELBO change 50k): background polygenic signal is detectable

• Locus shows many nominally associated variants (diffuse signal)
• SuSiE returns very large credible sets (many variants absorbed as "sparse" effects)

Key thresholds / parameters:

• Prior W (ABF): 0.04 (source: Wakefield 2009, Am J Hum Genet)
• Credible set coverage: 95% (adjustable via --coverage)
• Max signals L: 10 (adjustable via --max-signals)
• Min purity (SuSiE/SuSiE-inf CS filter): 0.5 average pairwise LD r² within set
• Convergence tolerance: max |ΔPIP| 0.1 in this locus?"

Output Structure

output_directory/
├── report.md                    # Primary markdown report
├── fine_mapping.json            # Machine-readable PIPs + credible sets
├── figures/
│   ├── pip_locus_plot.png       # Per-variant PIP coloured by LD r²
│   ├── regional_association.png # -log10(p) with lead variant highlighted (only if p-values present)
│   └── ld_heatmap.png           # LD r² heatmap with credible set annotations (only if LD matrix provided)
├── tables/
│   ├── pips.tsv                 # rsid, chr, pos, pip, cs_membership
│   └── credible_sets.tsv        # cs_id, size, coverage, lead_rsid, variants
└── reproducibility/
    ├── commands.sh              # Exact command to reproduce
    └── environment.yml          # Package versions

Dependencies

Required:

• numpy >= 1.24 — array maths, LD matrix operations
• scipy >= 1.10 — statistical functions
• pandas >= 1.5 — sumstats parsing
• matplotlib >= 3.7 — locus plots

Safety

• Local-first: No data upload; all computation is on-machine
• Disclaimer: Every report includes the ClawBio medical disclaimer
• Audit trail: reproducibility/commands.sh logs exact inputs and parameters
• No hallucinated science: All parameters trace to cited papers; model outputs are probabilistic, not clinical diagnoses

Integration with Bio Orchestrator

Trigger conditions — the orchestrator routes here when:

• Query contains "fine-map", "finemapping", "credible set", "PIP", "posterior inclusion"
• File has columns: beta/z + se (looks like GWAS summary stats)
• Query mentions SuSiE, FINEMAP, CAVIAR, ABF, polyfun

Chaining partners — this skill connects with:

• gwas-lookup: look up the lead variant before fine-mapping to confirm locus context
• gwas-prs: fine-mapped causal variants can be used as a more precise PRS variant set
• vcf-annotator: annotate the credible set variants with functional consequences

Citations

• Wang et al. (2020) JRSS-B — SuSiE algorithm
• Wakefield (2009) Am J Hum Genet — Approximate Bayes Factors for GWAS
• Cui et al. (2024) Nature Genetics — SuSiE-inf: improving fine-mapping by modeling infinitesimal effects

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: ClawBio
• Source: ClawBio/ClawBio
• License: MIT
• Homepage: https://clawbio.github.io/ClawBio/

Install and usage instructions live in the source repository linked above.