AXOLOTL

AXOLOTL: an accurate method for detecting aberrant gene expression in rare diseases using co-expression constraints.

We propose a novel unsupervised method AXOLOTL to identify aberrant gene expression events in RNA expression matrix. The method is useful for rare disease diagnosis. AXOLOTL effectively addresses biological confounders by incorporating co-expression constraints. The manuscript is being submitted to peer review jounrnals Jan 2024.

prepare software enviroment

We recommend to run AXO in docker enviroments. Create two docker image enviroments as follows:

1. R enviroment named 'r4.2:jammy': install OUTRIDER in R-4.2 using Dockerfile.

   cd ubuntu22_r4_outrider
   docker build --tag r4.2:jammy .

2. python enviroment named 'py3' install OutSingle and AXOLOTL python enviroment using Dockerfile. They is mainly implemented in python, thus data analysis modules (numpy, pandas, etc.) are needed.

   cd ../py3_outsingle 
   docker build --tag py3 .

Example Usage

Demo cohort have 1000 genes x 36 samples. Input: A RNA-seq expression matrix /testdata/df_cts.txt. Output: A aberrant score matrix /test/df_cts.txt.

Run AXO on demo data as:

bash script/demo.sh

more information

You may also have interest on https://github.com/gagneurlab/OUTRIDER & https://github.com/esalkovic/outsingle.