Enrichment Score(GSEA) compute Tools for Lincs data
A brief description of the tools is given below. The tools implement the function of parallel compute ES in multi-core environment. The Matlab implementation of the tools is currently the most mature. I used the 1ktools(https://github.com/cmap/l1ktools) tools to parse the .gctx file which stored gene profile data defined by Lincs(CMap) based on HDF5 file format. However,the performance is not very satisfactory. The C implementation to achieve at least 30 times faster than Matlab. But 1ktools can not parse the .gctx file by C. So I used Matlab to parse the .gctx file、extract the gene profile sets and write to .txt file. C will read the file 、complete parallel computing ES and write out the result in binary file . There also is a Matlab Script to read the binary files and can be later analysis. I will update the tools as they become available.
- Matlab R2009a and above
Enter the "pathtool" command, click "Add with Subfolders...", and select the directory ParalES/matlab.
- [ESquick.m]: Compute single Enrichment Scores.
- [ESScore.m]: parallel compute Multiple pairs of ES.
- [getSampleforMat.m]: parse the .gctx file and extract the gene profile sets we need.
- [ParalES.m]：main entrance . Setting Parameters and complete parallel Computing tasks.
- [getSample.m]: parse the .gctx file and extract the gene profile sets we need for Writting.
- [PreESforC.m] : Setting Parameters 、 extract the gene profile sets and write to .txt file.
- [importES.m] : Read the Enrichment Scores Result Matrix computed by C.
- [ParalES.c] read the .txt file 、complete parallel computing ES and write out the result in binary file
- [compileParalES.sh]: compile the C source code to Executable files.
- [runParalESLinux.sh]: run Executable files in Linux.
The CMAP Cloud API offers programmatic access to annotations and perturbational signatures in the LINCS L1000 dataset via a collection of HTTP-based RESTful web services. You can get the .gctx file stored gene profile data by the Website.