PhasingReads

Phasing Reads

This code details how bulk ATAC-seq, RNA-seq, and WGBS data were phased using the maternal and paternal genomes of the individuals under study. The subdirectories include the scripts for each different aligner (BWA-mem for ATAC, STAR for RNA, and Bismark for WGBS) and the Python scripts used to phase the outputs of these aligners.

The general workflow is detailed below:

---

Create a combined Genome file of maternal and paternal genome by renaming all chromosomes to be either maternal or paternal, and then cat the files together.

sed 's/^>chr./&_maternal/' HG00741_mat_ragtag.fasta > HG00741_mat_ragtag_marked.fasta
sed 's/^>chr./&_paternal/' HG00741_pat_ragtag.fasta > HG00741_pat_ragtag_marked.fasta
cat HG00741_mat_ragtag_marked.fasta HG00741_pat_ragtag_marked.fasta > HG00741_combined.fasta

Then align reads to this combined genome using BWA-MEM for ATAC-seq:

sbatch run_BWA_index.sh HG00741_combined.fasta\

Parameters.txt: THREADS REFERENCE READ1 READ2 OUTPUT

run_Batch_BWAMEM.sh Parameters.txt

STAR for RNA-seq:

sbatch /scratch/twlab/tekkey/PhasingData/scripts/run_STARAlign.sh $genomeDir $genomeFasta $GTFFile $read1 $read2

Bismark for WGBS:

Run snakemake pipeline in ReferenceAlignmentBias/PhasingReads/WGBS

For all three technologies, compile a parameter file with columns:

ATAC/RNA-seq:

FastQR1 FastQR2 CombinedBAM MatBAM PatBAM mat1Output pat1Output

WGBS:

FastQR1 FastQR2 CombinedBAM mat1Output pat1Output

Then run the following BASH scripts which call the corresponding python scripts

sbatch --array=1-#%# run_PartitionReadsATAC.sh ParamFileATAC.tsv
sbatch --array=1-#%# run_PartitionReadsRNA.sh ParamFileRNA.tsv
sbatch --array=1-#%# run_PartitionReadsWGBS.sh ParamFileWGBS.tsv \

Finally, get the partitioned reads for every individual split between maternal and paternal as fastq:
Make a param file with columns:

read1Fastq read2Fastq ParentalOuputFromPartition outputFastqR1 outputFastqR2 filePrefixForSorting

Then run:

sbatch --array=1-#%# SplitReads.sh Parameters.tsv

The result of which is 2 files per line, which are FastQs containing the reads associated with either the maternal or paternal genome. These are then realigned to hg38 and the parental genome using the same aligners as above.