Bonferroni Correction

[jonathan-bravo]

ExpansionHunterDenovo/scripts/core/common.py

Lines 125 to 128 in 998fd79

	def correct_pvalues(count_table):
	num_tests = len(count_table)
	for row in count_table:
	row["bonf_pvalue"] = min(row["pvalue"] * num_tests, 1.0)

From what I am reading the Bonferroni correction is the pvalue divided by the number of tests, but in the above function it looks like the pvalue is being multiplied by the number of tests. In my use case this is leading to all the bonf_pvalue(s) to default to a minimum value of 1.0...

Please let me know if I am understanding something incorrectly.

Cheers,
Johnny

[jingydz]

I have found that you are correct, so we should use the p-values instead of the Bonferroni-corrected p-values, right?

[jonathan-bravo]

If it were me, yes I would just use the p-values. Maybe you can output the number of tests and then generate the value yourself because when I attempted to correct the code I was getting nil values returned.