Step 2 — Motif Search

Scans every promoter for transcription-factor binding motifs and computes a hypergeometric over-representation p-value per motif.

Inputs

• Target FASTA — usually the promoters.fa from Step 1.

• Motifs — any combination of:

  • Free-text IUPAC consensus (one per line, NAME SEQ)

  • MEME file

  • Live import from PlantTFDB (157 species), AnimalTFDB (vertebrates + insects), JASPAR 2024, or HOCOMOCO v11.

Statistics

For each motif:

\[p = P(X \geq k) \;\text{where}\; X \sim \text{Hypergeom}(N, K, n)\]

with

\(N\)	Total number of promoters
\(K\)	Number of promoters in which the motif occurs at least once
\(n\)	Number of query promoters (e.g. a K-means cluster from Step 6)
\(k\)	Number of query promoters with a hit

Multiple-testing correction: Benjamini-Hochberg (cis_gs.enrichment.core.bh_fdr).

Outputs

• hits.csv — one row per gene × motif, with hit position, strand, raw and adjusted p-value.

• Significance Summary — collapsed table with one row per (gene × motif).

Gene-ID Resolution

Cis-GS adds three optional ID-mapping methods to bridge the common NCBI LOC### ↔ species-database mismatch:

1. Column swap — append XM_ / XP_ accessions to the exported CSV.

2. Mapping CSV — user-supplied two-column lookup.

3. GFF3 Dbxref expansion — pull every synonym from Dbxref= and locus_tag= attributes in the annotation.