Translating transcriptomics analysis into diagnostic workflows: clinical variant identification and interpretation in...

Area of Science:

• Genomics
• Transcriptomics
• Bioinformatics

Background:

• Genetic testing faces challenges in identifying causal variants and interpreting variants of uncertain significance (VUS).
• RNA sequencing (RNA-seq) shows promise in enhancing diagnostic capabilities for genetic disorders.
• Blood and fibroblast samples are crucial for comprehensive genetic analysis.

Purpose of the Study:

• To evaluate the diagnostic utility of RNA sequencing (RNA-seq) in patients with genetically undiagnosed diseases.
• To implement a machine learning algorithm for analyzing gene expression and splicing outliers.
• To explore both hypothesis-driven and hypothesis-free diagnostic strategies using RNA-seq.

Main Methods:

• RNA sequencing (RNA-seq) was performed on blood and fibroblast samples from 102 patients.
• A multi-modal machine learning algorithm, the Detection of RNA Outliers Pipeline (DROP), was used for outlier analysis.
• Gene expression and splicing patterns were analyzed to identify potential disease-causing variants.

Main Results:

• RNA-seq analysis contributed to the diagnosis of 22/102 (21.6%) patients.
• The approach successfully aided genetic diagnosis in 12 patients without a prior genetic candidate.
• Additional insights were gained for 4 patients with known findings and 6 patients with unestablished disease mechanisms.

Conclusions:

• Blood transcriptomics, analyzed via RNA-seq, holds significant clinical and scientific value for diagnosing genetic diseases.
• A framework for integrating RNA-seq into diagnostic guidelines (ACMG/AMP) is proposed, utilizing criteria like PVS1 and PP4.
• Further research is needed to refine thresholds and incorporate diverse RNA-seq data (e.g., nonsense-mediated decay, splicing completeness).