/sc-compReg_matlab

Primary LanguageMATLAB

sc-compReg_matlab

The comparison of gene regulatory networks between disease versus healthy or between two different treatments is an important scientific problem. We propose sc-compReg as a method for the single cell comparative regulatory analysis based on single cell gene expression (scRNA-seq) and single cell chromatin accessibility data (scATAC-seq) from two different conditions. Our software sc-compReg can be used as a stand-alone package that provides joint clustering and embedding of the cells from both scRNA-seq and scATAC-seq, and construction of differential regulatory networks across two conditions.

Requirment

Matlab 2018 or later version; Tested on Linux; The core functions (exclude pre-processing) are also tested on Windowns too. The runing time on the example data (full data) is about 30 minuates on a "normal" desktop computer.

Run sc-compReg:

Run sc-compReg by following two steps:

Step 1: get cell label (here we recomend CoupledNMF method)

Please see run.sh under CoupledNMF_2.0 folder

Step 2: comparative Regulatory analysis

1) pre-processing: 
     edit genome version in line 4 of processing_data_sc_compReg.sh, we support hg19, hg39, mm9 and mm10
     provide Peak name file of sample 1 and sample 2. example: PeakName1.txt and PeakName2.txt
     bash processing_data_sc_compReg.sh
     This step may take long time becasue it scan motif across entire genome (to generate the TFbinding matrix used in next step).
2) run sc-compReg:
     See example in demo.m file. Files are in ./example/
     note 1: the input is the  sample1.mat file and sample2.mat file. Plesae see example.
     note 2: Generatation of TFbinding.mat file is time consuming, so we have provided this data in this example. To generate your TFbinding on your owen data, please use 3 line codes start with %:
        load('./prior/MotifMatch_human_rmdup.mat')
        TFName=intersect(Symbol,unique(Match2(:,2)));
        TF_binding=mfbs(TFName,Element_name,motifName,motifWeight,Match2);

output

     Map: linked subpopulation, each row represent one linked subpopulation. First number is the cluster index in sample 1 and the second number represent the cluster index in sample 2.
     diffNet: cell array diffNet{1,i} represent the diffrential regulatory network of i-th linked subpopulation.