LUMPY and CNVnator

lumpy_cnvnator.sh

WIN=100
SAMPLE="NA12878"
SAMPLE_BAM="NA12878_S1.bam"

cnvnator -root $SAMPLE.$WIN.root -genome GRCh37 -tree $SAMPLE_BAM
cnvnator -genome GRCh37 -root $SAMPLE.$WIN.root -his $WIN -d /shared/genomes/b37/full/chroms
cnvnator -root $SAMPLE.$WIN.root -stat $WIN
cnvnator -root $SAMPLE.$WIN.root -partition $WIN
cnvnator -root $SAMPLE.$WIN.root -call $WIN > $SAMPLE.$WIN.cnvcalls.txt

~/src/lumpy-sv/scripts/cnvanator_to_bedpes.py \
    -c NA12878.$WIN.cnvcalls.txt \
    -b 600 \
    --del_o del.$WIN.bedpe \
    --dup_o dup.$WIN.bedpe

#### DEPENDING ON IF YOU BAMS ARE chr1 or 1 YOU MAY NOT NEED THIS STEP
cat del.$WIN.bedpe | sed -e "s/chr//g" > del.$WIN.nochr.bedpe
cat dup.$WIN.bedpe | sed -e "s/chr//g" > dup.$WIN.nochr.bedpe

~/src/lumpy-sv/scripts/bedpe_sort.py \
    -b del.$WIN.nochr.bedpe \
    -g ~/scratch/cnvnator/genome.txt\
    > del.$WIN.nochr.posSorted.bedpe

~/src/lumpy-sv/scripts/bedpe_sort.py \
    -b dup.$WIN.nochr.bedpe \
    -g ~/scratch/cnvnator/genome.txt\
    > dup.$WIN.nochr.posSorted.bedpe

~/src/lumpy-sv/bin/lumpy \
    -mw 4 \ 
    -tt 1.0 \ 
    -pe bam_file:$PEBAM,histo_file:$HISTO,mean:$MEAN,stdev:$STD,read_length:100,min_non_overlap:100,discordant_z:$Z,back_distance:20,weight:1,id:PE,min_mapping_threshold:10 \
    -sr bam_file:$SRBAM,back_distance:20,min_mapping_threshold:10,weight:1,id:SR,min_clip:20 \
    -bedpe bedpe_file:dup.$WIN.posSorted.bedpe,weight:3,id:DUP \
    -bedpe bedpe_file:del.$WIN.posSorted.bedpe,weight:3,id:DEL

Check these out: https://github.com/arq5x/bedtools2/tree/master/genomes

…

On Dec 18, 2020, at 7:59 AM, Nitin Ravikanthachari ***@***.***> wrote: ***@***.*** commented on this gist. What format should the genome file be in for the bedpe_sort.py script? I tried using the reference genome in fasta format and it returned a blank file. ./bedpe_sort.py -b del.sample.nochr.bedpe -g reference.fasta > del.sample.nochr.posSorted.bedpe — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.