WimSpee
commented
4 years ago I used fgbio AnnotateBamWithUmis to add the UMI tags to the BAM files that bwa-mem created on the R1 and R2 fastq files. http://fulcrumgenomics.github.io/fgbio/tools/latest/AnnotateBamWithUmis.html
Then marked duplicates using Picard. See the script below.
Not sure if this is also possible in bcbio. If so how. If not, if there are alternatives ways to do the same in bcbio. If not, if this is maybe something to add to bcbio.
After this script the BAMs can go in to a standard variant calling pipeline.
# Get a list of R1 fastq file and then a list of sample names
find ./ -name "*.1.fastq.gz" > R1.fastq.list
cut -d "/" -f 2 R1.fastq.list | cut -d "." -f 1 > R1.fastq.list.samples
# Loop over the sample names to create the sam files
while read sample; do bwa mem my_ref.fa ${sample}.1.fastq.gz ${sample}.2.fastq.gz -t 20 -R "@RG\tID:${sample}\tSM:${sample}" > ${sample}.sam ; done < R1.fastq.list.samples
# Loop over the sample names to create the bam files
while read sample; do samtools view -b ${sample}.sam > ${sample}.bam ; done < R1.fastq.list.samples
# Loop over the sample names to create the sorted bam files
while read sample; do samtools sort ${sample}.sam -o ${sample}.sorted.bam ; done < R1.fastq.list.samples
# Loop over the sample names to create the bam files that have the UMI tags for each read
while read sample; do fgbio AnnotateBamWithUmis -i ${sample}.sorted.bam -f ${sample}.index_1.fastq.gz-o ${sample}.sorted.umi.bam METRICS_FILE=${sample}.sorted.umi.metrics.txt BARCODE_TAG=RX ; done < R1.fastq.list.samples
# Loop over the sample names to create the bam files that have duplicate reads marked
while read sample; do picard MarkDuplicates INPUT=${sample}.sorted.umi.bam OUTPUT=${sample}.sorted.umi.dedup.bam METRICS_FILE=${sample}.sorted.umi.dedup.metrics.txt BARCODE_TAG=RX ; done < R1.fastq.list.samples
# Loop over the sample names to create the indexes for the deduplicated bam files
while read sample; do samtools index ${sample}.sorted.umi.dedup.bam ; done < R1.fastq.list.samples
naumenko-sa
Version info
Additional context I have some targeted sequencing data where after de-multiplexing I have 3 fastq files per sample.
In the third fastq file there is a 12 bp UMI for each read pair in the other fastq files.
Is it possible to align and variant call these using bcbio? How should the fastq sample table or final YAML file look?
I could not find this information in documentation, but maybe I did not look in the correct place.
Thank you.