LTLA / PseudoCount2018

Some thoughts on choosing an appropriate pseudo-count for log-transformation of scRNA-seq counts.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Errors upon log-transformation of scRNA-seq data

This repository contains code and manuscript files for the paper Overcoming systematic errors caused by log-transformation of normalized single-cell RNA sequencing data by Lun (2018).

The report subdirectory contains LaTeX files for the report. The scripts subdirectory contains R code to reproduce the analyses described in the report:

  • spurious_sim.R, to examine the non-zero log-fold change induced by the log-transformation.
  • substructure.R, to examine the introduction of structural artifacts due the log-transformation.
  • alternatives.R, to examine the performance of other transformations.
  • check_ercc.R, to examine the behaviour of the log-transformation on real data.
  • power_sim.R, to examine the consequences on detection power with an increased pseudo-count.

About

Some thoughts on choosing an appropriate pseudo-count for log-transformation of scRNA-seq counts.


Languages

Language:TeX 76.0%Language:R 24.0%