yixf-self / hic-pipeline Goto Github PK

The pipeline performs Fit-hi-c analysis with fixed size windows. The pipeline consists of two stages: mapping_all and hic_pipeline.
 

In mapping_all:

runall: This script calls all the other steps below in mapping/filtering pipeline
pipeline: This script defines all the variables and a step to be run with runall
step1_align: Align the reads using BWA to the reference genome. You will need the appropriate BWA index file.
step2_sort: Filter for number of mismatches, mapping quality and unique mapping. 
step3_joinAllCombinations: Take a join of the separately mapped XYZ_1.fq.gz and XYZ_2.fq.gz files to get paired reads.
step3_removePCRdups: Remove the read pairs that map to exactly same coordinates (both ends). Keep only one pair out of many copies.



In hic-pipeline: 

Assignment to fixed size windowing. Process genome for mappability. Then, the normalization using ICE and applying Fit-Hi-C.
This pipeline assumes that there are read pairs under the cleanedPairs folders.

    In pipeline:
    runall: This script calls all the other steps below
    step1 in runall: create all the fixed size fragements. Step1 needs to be run only once for a given resolution/window size. Runs intractively 

    In bin:
    step2_get-contactCounts-atFixedWindowSize: get contacts at fixed window size
    step3_preprocess_for_bias_pipeline.sh: preprocess the genome to get mappability per window. Note: needs to be run only once for a particular resolution.
    step4_sparseICEnormalize: normalizes each library using sparse ICE implementation amd calculate biases for each window. 
    step5-ContactsPerFragment: calculate total numbmer of contacts coming out of each window
    step6_fithic: perform fit-hic on data before and after ICE.
 
    In data/reference_genomes/hg19: chromosome length and HindIII fragmens files. 

     
In hic-pipeline_repl_combined:

The same as hic-pipeline, but combines replicates in step2. Additional paramenters need to be specified in step 2.