API Documentation¶

epitome Module¶

epitome is a computational model for predicting ChIP-seq peaks in a new cell type from chromatin accessibility and known ChIP-seq peaks from ENCODE. This module also includes scripts for processing ENCODE peaks.

Models¶

`PeakModel`(dataset[, test_celltypes, …])	Model for learning from ChIP-seq peaks.
`EpitomeModel`(args, *kwargs)

Helper functions¶

`download_and_unzip`(url, dst)	Downloads a url to local destination, unzips it and deletes zip.
`bed2Pyranges`(bed_file)	Loads bed file in as a pyranges object.
`indices_for_weighted_resample`(data, n, …)	Selects n rows from data that have the greatest number of labels (can be weighted) Returns indices to these rows.
`get_radius_indices`(radii, r, i, max_index)	Gets indices for a given radius r in both directions from index i.

Vizualization functions¶

`joint_plot`(dict_model1, dict_model2[, …])	Returns seaborn joint plot of two models.
`plot_assay_heatmap`(matrix, cellmap, assaymap)	Plots a matrix of available assays from available cells.
`heatmap_aggreement_from_model_weights`(model)	Plots seaborn heatmap for DNase weights of first layer in network.
`calibration_plot`(truth, preds, assay_dict, …)	Creates an xy scatter plot for predicted probability vs true probability.

Constants¶

Dataset

Enumeration determining train, valid, test or runtime.

EpitomeDataset¶

EpitomeDataset([data_dir, targets, cells, …])

Dataset for holding Epitome data.

Data Generator Functions¶

`load_data`(data, label_cell_types, …[, …])	Takes Deepsea data and calculates distance metrics from cell types whose locations are specified by label_cell_indices, and the other cell types in the set.
`generator_to_tf_dataset`(g, batch_size, …)	Generates a tensorflow dataset from a data generator.

Conversion¶

RegionConversion(base_bed, compare_bed)

Class for dealing with genomic region conversions.