new_majiq.PsiCoverage.from_sj_lsvs

classmethod PsiCoverage.from_sj_lsvs(sj, lsvs, minreads=10.0, minbins=3.0, num_bootstraps=30, pvalue_threshold=1e-07)

Create PsiCoverage from SJExperiment and Events.

Parameters:

  • sj (SJExperiment) – Intron and junction coverage from an experiment

  • lsvs (Events) – Events over which PsiCoverage will be defined

  • minreads, minbins (float) – Quantifiability thresholds

  • num_bootstraps (int) – The number of bootstrap replicates for bootstrapped estimates

  • pvalue_threshold (float) – P-value threshold for removing stacks under leave-one-out Poisson model of per-bin read coverage, for both raw and bootstrapped coverage (Set to nonpositive value to skip stack detection)

Returns:

rna_majiq.PsiCoverage

Notes

The pvalue_threshold for stack removal is applied to both raw and bootstrapped coverage. This differs from the behavior in MAJIQ v2, where stack removal was only applied to bootstrapped coverage. In this sense “raw” coverage is only after stack detection.