Per-locus features

uchrom.fea computes what describes a locus across many cells: distance maps and contact frequencies, the variance of positions along each axis (what the callers test), signals summed over intervals, peaks of a signal, annotation and sequence content — and projects interval features onto the locus axis (bin_tracks) so that they sit next to imaging or sequencing signals. Every feature is recorded in cd.uns["feature_registry"].

Feature

API

distance maps (exact, streaming), mean distances, contact frequency, radius of gyration

distance_map, DistanceMap, mean_distance_matrix, contact_frequency, radius_of_gyration

axis-wise variance (the callers’ statistic)

axis_variance_cube, axis_weight, filter_normalize

signals over intervals; intervals onto bins

aggregate_tracks_by_interval, project_interval_features_to_bins

peaks of a signal (MACS-style)

call_peaks_from_track, call_macs_bdgpeaks_from_signal, add_peak_features

annotation (GTF) and sequence (FASTA) features

add_annotation_features, add_sequence_features

Tutorials

API: uchrom.fea.