evaluate¶

crystal.evaluate contains functions to evaluate methods for modeling clusters.

crystal.evaluate.evaluate(bed_iter, regions, sizes=None, label='', ax=None, **plot_kwargs)[source]¶

Evaluate the accuracy of a method (model_fn) by comparing how many fall into a set of “truth” regions vs. outside.

Parameters:	bed_iter : iterable iterable with each element of (chrom, start, end, p-value, n_sites) can be regions or features. regions : file BED file of regions. sizes : tuple(int,) which size regions to test. default (None) is all sizes. label : str label for plot legend ax : axis

crystal.evaluate.plot_roc(ax, r, plot_kwargs={})[source]¶

Plot ROC for a given result.

Parameters:	ax: matplotlib axis r: dict return value from evaluate_method()

crystal.evaluate.plot_times(ax, results, labels, colors=None)[source]¶

Plot the times taken for each result in results from evalute_method()

Parameters:	ax: matplotlib axis results: list of list of dict list of return values from model_clusters() colors: list list of colors for matplotlib

crystal.evaluate.write_modeled_regions(modeled_clusters, p_cutoff, out_fh)[source]¶

Write a region bed file suitable for use in evaluate_modeled_regions().

Parameters:	modeled_clusters : list output from model_clusters() p_cutoff : float values < this are set as true out_fh : filehandle where to write the data

crystal.evaluate.write_region_bed(feature_iter, true_regions, out_fh)[source]¶

Write a region bed file suitable for use in evaluate_modeled_regions(). given true regions (likely from an external program, otherwise use write_modeled_regions()).

Parameters:	feature_iter : iterable of Features true_regions : file BED file containing true regions out_fh : filehandle where to write the data