Alternative to K-means clustering in infer_trajectory()

[alikhuseynov]

Hi there,
again, great tool!

Few questions:

• I wonder if one can use an alternative clustering approach, instead of k-means? is this issue #3 relevant here? just want to clarify :)
• ..if one can somehow supply infer_trajectory() with already existing clusters and see how it would predict the trajectory?
• Also it looks like one has to provide eg set.seed(123) before running infer_trajectory() since k-means starts randomly.
• what would you suggest on extract_modules() for large dataset? issue #32

Many Thanks
A.