search

KrishnaswamyLab / MAGIC

MAGIC (Markov Affinity-based Graph Imputation of Cells), is a method for imputing missing values restoring structure of large biological datasets.
GNU General Public License v2.0
341 stars 97 forks source link

magic() documentation in R #119

Closed ronammar closed 6 years ago

ronammar commented 6 years ago

In Rmagic::magic() the documentation reports that the default number of principal components is 20, but the default in magic() is set to 100.

scottgigante commented 6 years ago

Hi @ronammar, thanks for the report. We will update the documentation in the next release. From our experience, 100 principal components is sufficient for most datasets and provides reasonable speed of imputation.

Thanks again!

ronammar commented 6 years ago

Great - thanks! I figured it doesn't really matter based on your published methods:

We also note that MAGIC is relatively robust to the number of principle components selected, within a reasonable range

But would remove ambiguity as to what the default should be.

scottgigante commented 6 years ago

Thanks again! The fix has been added to the development branch and will be included in the next release.