Long-read transcriptome analysis using IsoRanker for identifying pathogenic variants in Mendelian conditions.

Variant prioritisation from long-read transcriptomics: outlier gene and isoform expression, allelic imbalance and nonsense-mediated decay revealed by cycloheximide treatment.

IsoRanker is a long-read transcriptome sequencing framework that prioritises functionally relevant variants by detecting genes and isoforms with outlier expression, allelic imbalance and/or nonsense-mediated decay. The authors generated paired cycloheximide-treated and untreated fibroblast transcriptomes from 31 individuals — 3 with known transcript-altering rare variants and 28 with unsolved conditions — linked to phased long-read genomes. IsoRanker recovered the known transcript alterations, and exploratory subsampling suggested that prioritisation was largely preserved down to cohorts of 11 individuals and about 5 million full-length transcripts per individual, with performance depending on the choice of de novo isoform caller, particularly for NMD-sensitive isoforms. Among the 28 unsolved cases, IsoRanker deprioritised 8 of 10 fibroblast-expressed candidate splice-site variants while nominating 4 new leads; in one individual it prioritised HARS1, revealing bi-allelic non-coding variants that together produced a partial HARS1 loss of function and informed targeted therapy.

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

The ability to deprioritise 8 of 10 splice candidates is worth as much as the new leads: in clinic, it is these variants of uncertain significance that clutter files and durably worry families. Sensitivity to the choice of isoform caller is the real obstacle to routine use — two laboratories with two different callers will not produce the same candidate list, making prior harmonisation essential. The tissue limitation also stands: fibroblasts only inform on the genes they express, which excludes a large share of neurological and muscular conditions from the outset.

Analysis by Dr Thibaut Benquey