Likelihood-based calibration improves the clinical utility of JAG1 functional data for variant classification.

Conversion of scores from a multiplexed assay of variant effect in JAG1 into log-likelihood ratios of pathogenicity, translatable into strong, moderate or supporting ACMG/AMP evidence weights.

Multiplexed assays of variant effect (MAVEs) provide functional data at scale, but their readout must be translated into the language of clinical classification. Revisiting their MAVE of JAG1 variants, the primary cause of the autosomal dominant multisystemic Alagille syndrome, the authors calculated log-likelihood ratios of pathogenicity for each variant score, allowing direct conversion of functional data into ACMG/AMP evidence weights. Calibration improved separation of known benign and pathogenic variants and increased the number of classified abnormal missense variants from 486 to 610, binned into strong (n = 1), moderate (n = 340) and supporting (n = 269) evidence towards pathogenicity. Applied retrospectively to 29 individuals carrying a JAG1 variant of uncertain significance identified through clinical diagnostic sequencing, it yielded nine variants (31%) with abnormal data meeting supporting (n = 3) or moderate (n = 6) weight, of which six (21%) were upgraded to likely pathogenic or pathogenic. The authors argue these models apply to further MAVEs.

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

This is exactly what MAVEs lack: without calibration, a functional score remains raw data that the laboratory applies by judgement, with the risk of granting PS3 too generously. Six variants of uncertain significance reclassified out of 29 in a real diagnostic cohort is a credible yield, and the method transfers to any MAVE with enough reference variants — that, beyond Alagille syndrome, is the paper's real reach. Calibration nonetheless inherits the limits of the underlying assay: a variant not tested, or tested in an irrelevant cellular context, gains nothing.

Analysis by Dr Thibaut Benquey