Allow users to select number of peaks to keep instead of stringency thresholds.

[callum-b]

One setup I have used in the past is to call a set number of peaks for different experiments (for example, binding of a TF in various conditions) and then comparing how those peaks overlap on genomic features. An option in MSPC to keep the 50000 top peaks as determined by the combined p-value would allow this quite easily.

I know this can be done by setting a very permissible threshold and then filtering the output files, but if there is a way to do this during normal MSPC runtime, it would avoid sorting a massive bed file to find the top peaks.

[VJalili]

Agree, a parameter such as --limit could be helpful, similar to directives on SQL queries. I wonder what would be minimal and comprehensive design.

• One --limit argument that is applied to all the outputs MSPC generates.
• One --limit per output type.

The latter gives you finer level control, but it would convoluted CLI options. The former is limited, but covers most use-cases.