Ways to make MiXCR faster

[SirKuikka]

Hi,

I am running MiXCR with this command

mixcr analyze rna-seq --species hsa fastq1 fastq2 output

Are there ways to speed up the method without significantly compromising accuracy? I have a BAM file generated from read alignment using Dragen, and I attempted to accelerate the process by pre-selecting reads from the TCR and BCR loci. This approach did improve performance; however, I noticed a slight discrepancy in the number of clones identified when comparing the full dataset analysis to the pre-filtered one. Do you have any experience from pre-filtering the reads like this?

[mizraelson]

Hi, yes pre-selecting reads might help, but also increasing the CPU and memory resources allocated for the process.