Trim or not to trim in RNAseq based TCRs clonotypes identification

VERÓNICA BARONETTO; GUADALUPE NIBEYRO; JUAN CARLOS VAZQUEZ; ELMER FERNÁNDEZ

Conferencia; 2nd Women in Bioinformatics & Data Science LA Conference; 2021

Women in Bioinformatics & Data Science LA

T cell receptors (TCRs) are responsible for the identification of tumor neoantigens presented on the surface of the tumor cell, a process favored by checkpoints' inhibitors immunotherapy such as anti-PD-1, anti-PD-L1 or anti-CTL4, leading to an effective antitumor response. The TCRs, are generated through the process of VJ recombination at the Tcrα and V(D)J recombination at the Tcrβ loci through complex rearrangements by means of splicing, single nucleotide polymorphism, insertion and deletions events, that requires specifically designed alignment algorithms (i.e MiXCR, IMGT-V-QUEST) for their identification on sequence transcript data. Such approaches turn crucial for the understanding of the immune microenvironment-tumor cells association. The TCR identification starts with sequence raw data (FASTQ files), that should be aligned against a TCR loci reference sequence. Since sequence reads do not always havecorrect signals, being longer than the expected or containing adapters, or low quality reads, a process known as ?trimming? that removes adapters and low quality bases is often assumed as necessary. Due to its high computational cost, the role of this preprocessing step by trimgalore is compared against the use of raw data in the detection of TCR clonotypes in RNAseq samples by MiXCR platform to verify this use is truly necessary or worthless. A total of 833 clones (TCRα and TCRβ) were detected between both approaches with 95% of TCRs overlap. Only 2% were detected using trimmed sequences and only 3% using raw data. In 4 out of 9 of the trimmed samples a maximum of 4% of TCRs clones were exclusive meanwhile in 6 of 9 a maximum of 4% more clones were detected using raw data. Our results indicate that the use of ?trimming? pre-processing is not necessary for TCR clonotype identification, with the consequent impact on saving computational resources and reducing results delivery times.