cluster_chem_data: Run logistic singular value decomposition on the binary...

In blind-contours/ChemLogic: Pipeline of methods that take in chemical fingerprint and molecular physical information and creates interpretable rules based on this information which leads to nuclear receptor binding

Description

This function takes in a binary matrix, assuming that the variable names include the fingerprint name such as "Pubchem"

Usage

cluster_chem_data(receptor_type, data, rank = 3, filter_type = "Pubchem")

Arguments

receptor_type	Used for the kmeans plot from the SVD results, this is a characer string such as "Androgen"
data	matrix or dataframe that has all binary data except for possibly Name, Label_names, and Label, which are removed
rank	rank of the SVD that decomposes the binary molecular feature data before using the PCs in kmeans.

Value

Results from the logistic SVD

Plot of Kmeans for visual inspection before creating clusters

blind-contours/ChemLogic documentation built on July 7, 2020, 5:34 a.m.