Cis mapping¤

Cis mapping tests variants within a window around each molecular phenotype and reports one selected association per phenotype. The output includes the lead variant, its nominal association, and a gene-level adjusted p-value.

The --window value defaults to 500,000 bases. By default, the interval extends from TSS - window through TES + window. Add --tss-centered to instead use TSS - window through TSS + window.

Choose a calibration method¤

Use --spa --acat for fast score-test aggregation without permutations; SPA is strongly recommended because ACAT is sensitive to variant p-value calibration. Omit --acat to use Beta permutation, with --nperm controlling the number of shuffles. The methods need not produce the same p-values or discoveries. See Tests and gene-level calibration for the statistical tradeoffs.

The Quickstart contains complete commands for both methods.

Select regions and phenotypes¤

Use --chr LABEL to restrict the scan to an exact chromosome label shared by the inputs, and --gene-list PATH to select phenotype IDs from a file. Identifier lists contain one ID per line. Using a gene list with the full phenotype matrix preserves library-size offsets computed before gene selection; see Offsets when the input file itself has already been restricted.

Execution and fitting¤

Score and SPA scans reuse one null fit per phenotype; permutation scans fit a null model for each shuffle. Compiled kernels are reused across cis-window sizes. See Run large scans for compilation and memory behavior.

GLM fitting is controlled by --tol, --gtol, --max-iter, and --step-size. See Troubleshooting for defaults and stopping rules.

Inspect results¤

Follow Post-process cis results to filter failures and apply study-level FDR correction. The gene-level pvalue_adj is not an across-gene FDR value.

Genes with no variants in the requested window or no phenotype variance are skipped. See Cis output for the complete result contract.