Entropy on strand

[Ge0rges]

Hey @ArtRand,

I just wanted to write to ask for a feature request for entropy which would be to make it "strand aware" by allowing us to specify the strand for each region -, + or ..

It would also be convenient if the documentation specified the headers for the output files when --regions is specified.

Thanks!

[ArtRand]

Hello @Ge0rges,

I agree that the BED file should direct the strand to use. I'll be sure to add it along with the multi-base work.

[Ge0rges]

Also I've noticed sometimes the output regions.bed is empty with no error printed?

[ArtRand]

The final log will report 0 regions processed successfully in this case. There is always a bit of a balance to strike with respect to informing the user why something was ineligible to calculate a result and making the logs very verbose and hard to follow. Perhaps a better solution is to tabulate how many regions failed and their reasons?

[Ge0rges]

Hey @ArtRand could you let me know what schema for the output is when --regions is specified?

[ArtRand]

Hello @Ge0rges,

The schema is:

col	Name	Description	type
1	chrom	chromosome of the region	str
2	start	0-based start position of the region	int
3	end	0-based end position of the region	int
4	region_name	name of the region from the input BED file	str
5	mean_entropy	average entropy of the passing windows included in the region	float
6	strand	strand of the region {+, -, . }	str
7	median_entropy	median entropy of the passing windows included in the region	float
8	min_entropy	minimum passing window entropy	float
9	max_entropy	maximum passing window entropy	float
10	mean_num_reads	average number of reads used in the passing windows' entropy calculation	float
11	min_num_reads	minimum number of reads used in the passing windows' entropy calculation	int
12	max_num_reads	minimum number of reads used in the passing windows' entropy calculation	int
13	successful_window_count	number of passing windows in the region	int
14	failed_window_count	number of failed windows in the region	int

You can also pass the --header flag to get a header line in the output.

I'll add this to the documentation.