API reference

Everything is imported from the top-level package:

import universalmotif as um

The functions are grouped by what they do:

• Motifs: the Motif class, and creating, converting and modifying motifs.
• Reading and writing: motif files, and FASTA files of sequences.
• P-values and scores: thresholds, and the scores of individual matches.
• Scanning: finding motif hits in sequences.
• Comparison and clustering: similarity, merging and trees.
• Enrichment and discovery: enrichment tests and de novo motif discovery.
• Sequences: random and shuffled sequences, backgrounds, complexity and masking.
• Genomic intervals: matched backgrounds and proximity to anchors.
• Plotting: logos, trees and plots of results.
• Databases and Biopython: motifs from JASPAR, HOCOMOCO and CIS-BP, and conversion to and from Biopython.
• Deep learning: convolution filters, one-hot sequences and TF-MoDISco results.
• Example data: synthetic motifs, sequences and files for trying things out.

Most functions that do a lot of work take an nthreads argument, which is 1 by default or 0 to use every core. Functions that use random numbers take a seed, and the same seed gives the same results again; without one they draw a fresh seed each time.