search

Sage-Bionetworks / NF_LandscapePaper_2019

This repository hosts all the code used to generate analyses and figures for the landscape paper
3 stars 1 forks source link

Add PCA to Fig 1 analysis and create Fig 2 #62

Closed sgosline closed 4 years ago

sgosline commented 4 years ago

For further quality control we will plot samples in PCA plot.

I will recreate fig1 based on LVs and immune signatures.

sgosline commented 4 years ago

Ok. I've changed my mind on this. Hear me out - I think the clustering will always show batch effects, specifically the LVs are even worse because we know that some pick up batch effects.

So I think we should double down on the LVs/genes/proteins that specifically predict tumor type. They still might not 'cluster' by tumor type, because the RF is able to have non-linear logic. So clustering is not the best idea.

We should use RF performance to motivate the focus on LVs.

sgosline commented 4 years ago

So we can close this after #60 is approved, because I did the heatmap of the LVs.