How does Omics Playground handle missing data?

37 views

Skip to first unread message

BigOmics Analytics

unread,

May 19, 2021, 3:52:39 AM5/19/21

to Omics Playground

[email Thorben S. 19.05.2021]

"I would like to know how the OmicsPlayground handles missing values. Are they imputed in any kind or left as missing and not conisdered for the PCA and clustering?"

BigOmics Analytics

unread,

May 19, 2021, 4:21:37 AM5/19/21

to Omics Playground

Hi Thorben,

At the moment missing values are set to 0 (zero). The Omics Playground uses the function pgx.createPGX in the file pgx-compute.R

Note that only real NA (or empty) values in the counts file are supposed to be "missing". Zero values are assumed to be real zeros. If zero-value imputaton is not what you want, and you want to impute your "missing" values differently, (currently) you must do that manually before uploading the CSV file.

If you would like to see some imputation methods or automatic imputation implemented in Omics Playground, please submit a feature request and we may implement such a feature if many appreciate that.