Improvement Request: downsample to the maximum of either the file event count or target

[tjchen]

in downsample.R please give the option of downsampling to the maximum of either the file event count or target. Currently it is a hard threshold on the event count and fails if the number of events in the file are less than the downsampling event number.

[esimonds]

Good idea. While we're mucking around I there, is support for "percent" of
cell events currently supported in Driver.R? (note that I'm not talking
about "percentile of density"). It's the default on Cytobank SPADE but I
suspect it was an improvement on their end.

On Saturday, March 1, 2014, TJ Chen [email protected] wrote:

in downsample.R please give the option of downsampling to the maximum of
either the file event count or target. Currently it is a hard threshold on
the event count and fails if the number of events in the file are less than
the downsampling event number.

Reply to this email directly or view it on GitHubhttps://github.com//issues/89
.

[zbjornson]

@esimonds yep, it's there:

(... in driver.r...)
downsampling_target_number=NULL,
downsampling_target_pctile=NULL,
downsampling_target_percent=0.1,
(...)