NanoTS: a deep learning tool for accurate SNP calling in nanopore long-read transcriptome data.

Deep learning-based tool detecting single nucleotide polymorphisms from nanopore transcriptome sequencing data, both direct RNA and cDNA

Accurate variant detection from nanopore long-read transcriptome data remains challenging. NanoTS is a deep learning-based tool that detects single nucleotide polymorphisms (SNPs) in diverse types of nanopore transcriptome sequencing data. The authors report that it outperforms existing methods, with F1 scores above 0.980 on direct RNA and above 0.966 on cDNA sequencing data, for SNPs supported by at least five reads. It shows strong improvements for allelically imbalanced variants. Finally, the authors show that it enables accurate detection and genotype calling of pathogenic variants underlying Mendelian disorders, highlighting its potential clinical utility.

Synthesis written by Geno'X. For the full original abstract, please refer to the source publication.

F1 scores above 0.980 on direct RNA and above 0.966 on cDNA, with a clear gain on allelically imbalanced variants, make NanoTS a serious candidate for SNP calling directly on the nanopore transcriptome. These figures apply, however, to SNPs supported by at least five reads and say nothing about positions covered by fewer; the demonstration on pathogenic Mendelian variants is framed as clinical potential, with no sample size or sensitivity in the abstract. Its diagnostic added value over variant calling on DNA remains to be established.

Analysis by Dr Thibaut Benquey