Co-expression analysis in Go

An implementation of weighted gene co-expression network analysis in Go. This is a final project for CMU's Programming for Scientists class.

Authors

• Serena Abraham - matrices.go
• Jon Luo - clustering.go
• Hanxi Xiao - io.go

All authors contributed to each set.

Example data from https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE115522

Getting Started

Prerequisites

You will need ~10GB free RAM.

Gonum is required to run the project.

go get -u gonum.org/v1/gonum/...

Unzip data.zip before running.

Install and run

go build
unzip data.zip
./coexpression

You will be prompted to choose testing options. We recommend Pearsons and Signed correlation network.

---

CO-EXPRESSION GRAPHS AND THEIR ANALYSIS

For the covariance tests, please enter P for Pearsons (faster) or B for BiWeightedCorrelation:

Enter S to build a Signed correlation network or U for Unsigned:

Gene clusters will be saved to clusters.txt with one cluster per line.

Approximate runtime on a Ryzen 3700X 8-core/16-thread CPU is 15 minutes.