Running umi_tools with barcodes and umi in different fastq files

[nbbarrientos]

Hi,

Not sure if this issue has been addressed in a different thread. If so, I'd appreciate someone directing me in that direction.

But if not, my question is whether umi_tools extract can be run under the scenario where my read1 and read2 files do not have any barcode or umi information but such information is in different fastq files. For example, read1 is in read1_1.fastq, read2 in read2_2.fastq, umi in umi_3.fastq, and barcodes in barcode_4.fastq.

I'd appreciate any help. Thank you.

[pterzian]

Hi,

This is how I am handling the case of UMI barcodes being in a third fastq file (not sure how to handle sample barcodes tho) :

fgbio -Xmx16g AnnotateBamWithUmis \
            -i "$bam" \
            -f "$r3_fastq" \
            -o "$bam_with_umi" \
            -t RX 

This way you skip the umi-tools extract command and only add the UMI barcodes to the header of reads in the bam file. You can then call the umi-tools dedup command on this annotated bam.

In my case I also had to add the -Xmx16g parameter to raise the java heap space from 2GB default allocated RAM to 16GB.

P.