Single-cell Preprocess: Add TF-IDF

[mstrazar]

An alternative to log(CPM+1) transformation of count data is the TF-IDF transform, adopted from text analysis. Similar to finding characteristic words describing a topic in the document, TF-IDF can be used to find stand-out genes ("terms") for each cell ("document").

It should be relatively straightforward to include this approach into Single-cell preprocess.

See https://bmcgenomics.biomedcentral.com/articles/10.1186/s12864-018-4922-4