suwonglab / timereg

time course regulatory analysis from paired gene expression and chromatin accessibility time course data

TimeReg

time course regulatory analysis from paired gene expression and chromatin accessibility time course data

At each time point we use the expression and accessibility data to define two types of scores for regulatory relations. The trans-regulatory score (TRS) quantifies the regulatory strength of a transcription factor (TF) on a target gene (TG), while the cis-regulatory score (CRS) quantifies the regulatory strengths of a regulatory element (RE) on a target gene. Based on these scores, we use Non-negative matrix factorization to extract the core regulatory modules that characterize different biological processes and/or subpopulations of cells, and we identify diver TFs (i.e. TFs driving expression changes between adjacent time points) based on changes in TFs expression and changes in CRS on REs with increasing accessibility.

Quickly start:

wget https://github.com/SUwonglab/TimeReg/archive/master.zip

unzip master.zip

cd TimeReg-master/

bash install.sh

Example data (RA inducntion of mESC)

bash tcPECA.sh

Run TimeReg

Edit TCRA_config file as shown un RA induction example and run following script:

cat Your_config_file TCRA.m > run_TCRA.m

matlab -nodisplay -nosplash -nodesktop -r "run_TCRA; exit"

timereg's Issues

link http://web.stanford.edu/~zduren/PECA/Thresholding-Based%20SVD_files/Prior_corr.tar.gz is invalid

Dear all,

I'm trying to use your TimeReg to run my data, but the link http://web.stanford.edu/~zduren/PECA/Thresholding-Based%20SVD_files/Prior_corr.tar.gz is invalid.

Can you provide another link to download the priori data?

Thank you very much

Updates on `PECA` and related repositories

Hi there,

I've been exploring a bunch of repositories that either contain the PECA software or have PECA as a major dependency. I've found these repos across a few accounts:

1. PECA by SUwonglab
2. PECA by durenzn
3. PECA-1 by durenzn
4. TimeReg by SUwonglab
5. vPECA by AMSSwanglab

I've previously submitted issues on the SUwonglab/PECA issue tracker and made an attempt to get in touch with @durenzn directly, both via "@" and by email. Unfortunately, those issues are still awaiting responses.

To get a better picture of the current state of PECA and cast a wide net, I'm posting this identical issue across all these repositories. Is PECA archived, or is it still under development or maintenance? If it's still being developed or maintained, where's that mainly happening?

Any information or guidance would be super helpful. Thanks!

How to set lambda?

Dear all,

I'm trying to use TimReg on my data, but I don't know how to give an appropriate lambda value for my data?

Can you give me more suggestions about how to pick an appropriate lambda value for my data?

Thanks for your great help!

how to caculate the K module?

Thanks for your software.

According to your Genome Research paper, I know that you caculated the K modules (subpopulation) using the script in "TFTG_co_module.m" and the TRS matrix. However, in your TimeReg manual, there is no item to tell us how to use this script. I'm a layman on matlab. Can you illustrate how to get the K using your script?

Error，In run_TCRA (line 40) Matrix dimensions must agree.

Thanks for your software. When I run the command, it report the error:

In run_TCRA (line 40)
Matrix dimensions must agree.

Error in run_TCRA (line 58)
Exp2=log2(1+Exp)-repmat(prior.Exp_median(f2(d2)),1,m)

=========================================

I have 9 time points data:
DorRe0min
DorRe30min
DorRe60min
DorRe90min
DorRe2hr
DorRe3hr
DorRe4hr
DorRe6hr
NormalCycle

In exampleData file, the rows of each file are:
716 DorRe0min_TRS.txt
716 DorRe2hr_TRS.txt
716 DorRe30min_TRS.txt
716 DorRe3hr_TRS.txt
716 DorRe4hr_TRS.txt
716 DorRe60min_TRS.txt
716 DorRe6hr_TRS.txt
716 DorRe90min_TRS.txt
17719 Exp.txt
716 NormalCycle_TRS.txt
9 SampleName.txt
715 TFName.txt
17719 TGName.txt

The column of these files are；
cat *txt | awk "{print NF}" | uniq
17720
9
17720
1

The config file is:
PECA_Module_dir = './.';
Expfile='./exampleData/Exp.txt';
sample_trs_files='./exampleData/sample_trs_info1';
TFName_file='./exampleData/TFName.txt';
TGName_file='./exampleData/TGName.txt';
Outdir='./exampleData/out';
pre_process_required=0;
species = 'human';
lambda=0.2;
K=[1,1,1,1,1,1,1,1,1];

%input format (tab delimited), here T is the number of time points
%1, Expfile: name (location) of the expression file. It is a T+1 columns file, first column is the gene symbol, 2 to T+1 columns are expression value (FPKM or TPM),
%2, sample_trs_files: T rows 2 column file, first column is the sample name (no space allowed), second column is the name of the TRS file from PECA2 (no space allowed).
%example: mESC ./PECA2/Results/mESC/TFTG_score.txt
%3, TFName_file: file name of the TF Names from PECA2
%4, TGName_file: file name of the TG Names from PECA2
%5, out_folder: name of output folder
%6, pre_process_required: 0 or 1. If you require preprocess, set this variavble 1, othervise 0. pre-process will remove some genes with constant expression.
%7, species: 'mouse' or 'human'.
%8, lambda: weight of correlation from your time course data (lambda) verse public data (1-lambda), it is continious value between 0 and 1.
%9, K: number of subpopulation on each time point, which is a 1*T vector.

=============================================

I wonder I set the wrong K parameter. I don't understand why your set the K as [1,3,3,4,4} in your provided example data.