pipline_HINT_for_scATAC

0. put sbatch_param to your local execution path (e.g. $HOME/.local/bin/)

sbatch_param enables you to set the parameters for sbatch command e.g. sbatch_param -f yourscript.sh -n yourjobname -a your1stparam -b your2ndparam -c your3rdparam where $aaaaa is the first parameter, $bbbbb is the second parameter, and $ccccc is the third parameter in your script

you can also run without sbatch_param directly using sbatch 02_helper.sh by setting all the parameters in the script

1. run script to split barcodes by clusters and samples

Rscript 01_getCluster.R

this will generate barcodes text file recording the cluster information in each sample named file.

2. run script to split bamfiles by the split barcode produced in step 1

02_all_run.sh will submit all jobs to slurm nodes by passing paramters to 02_helper.sh, where 02_helper.sh call split_bam_by_cluster.py to split bam files by clusters and samples

This will generate a lot of bam files, "$sampleName_$clusterName.bam" in the BAM folder

3. run HINT-ATAC on bam files

sh 03_hint_atac.sh includes a bunch of commands to run HINT-ATAC on bam files

01. merge all bam files that belong to the same cluster

02. call peaks for each cluster & perform footprinting

03. calculate the differential footprinting

4. Can use the merged peak to redo the counting of fragments count matrix

01. `Rscript 04_peak_count_matrix.R` first get the merged peaks from all clusters, then count the fragments using the union peaks