Information Theoretic Analysis of Gene Expression Data

Implements Surprisal analysis for gene expression data such as RNA-seq or microarray experiments. Surprisal analysis is an information-theoretic method that decomposes gene expression data into a baseline state and constraint-associated deviations, capturing coordinated gene expression patterns under different biological conditions. References: Kravchenko-Balasha N. et al. (2014) . Zadran S. et al. (2014) . Su Y. et al. (2019) . Bogaert K. A. et al. (2018) .

---

SurprisalAnalysis R package guidelines

🖥️ Installation

To install the R package:

install.packages('devtools')
devtools::install_github('AnniceNajafi/SurprisalAnalysis')

Usage

To use the R package you should follow the steps below:

I. Store gene expression data in a csv file with the first row holding the sample names and the first column holding the gene names.

II. Read the csv file and run the following code:

  input.data <- read.csv('expression_data.csv')
  results <- surprisal_analysis(input.data)
  

III. To run GO analysis on the patterns simply use the code below:

results[[2]]-> transcript_weights
percentile_GO <- 0.95 #change based on your preference
lambda_no <- 1 #change based on your preference
GO_analysis_surprisal_analysis(transcript_weights, percentile_GO, lambda_no, key_type = "SYMBOL", flip = FALSE, species.db.str =  "org.Hs.eg.db", top_GO_terms=15)

Use GUI from R package

Simply run the following code:

runSurprisalApp()

Web-based application

A web-based application based on the above has been deployed on this link.

Open source disclaimer

This is an open-source project based on a previously developed methodology. Requests or attempts on the expansion and further improvement of the code is welcome and encouraged.