The clinical Blended Genome Exome (cBGE) laboratory workflow utilizes the New England Biolabs NEBNext Ultra II FS DNA Library Preparation Kit and NEBNext Multiplex unique dual index adapter oligos set 1-4, which transformed genomic DNA extracted from blood or saliva into DNA libraries that are compatible with Illumina sequencing platforms. During the library construction workflow, gDNA was enzymatically fragmented, adapter-ligated, and barcoded to create a single PCR-free whole genome. Following PCR-free generation, an aliquot from the PCR-free whole genome library was taken through PCR amplification and exome selection. The PCR-free genome and whole exome were then rejoined and blended together pre-sequencing to deliver an optimal coverage balance and combined sequencing output. Library fragments were sequenced (2x150 base paired-end) using Sequencing-By-Synthesis (SBS) chemistry on the Illumina NovaSeq X Plus sequencer. Sequencing data were aligned to the GRCh38 assembly after discarding low-quality sequences. Illumina’s DRAGEN (Dynamic Read Analysis for GENomics) platform was used for demultiplexing, read mapping, genome alignment, read sorting, and duplicate marking. Sample quality metrics met the following thresholds: ≥ 90% of exome bases at or greater than 20X, genome mean coverage ≥ 1X, mean target coverage ≥ 60X, <2.5% contamination, ≥ 75% of reads mapped, and ≥ 95% of bases callable across the exome region. The DRAGEN pipeline generated a CRAM file, which was imputed using GLIMPSE2, a low-coverage whole-genome sequencing imputation tool. The imputed VCF was used as input for the PRSMix+ algorithms (see below) for the tested cardiovascular conditions.
The PRSMix+ framework optimizes the performance of PRS across multiple ancestries, reduces the risk of inappropriate reclassification as scores improve, and enables the ongoing improvement of the scores (Truong B et al., Cell Genom, Apr 2024). For each condition, all scores available from the PGS Catalog at the time of development are integrated into a single condition-specific score based on individual score performance in the US-based NIH All of Us Research Program. The PRSMix+ scores are adjusted by principal components of genetic similarity group to account for raw PRS differences due to ancestral background (Hao L et al., Nat Med, May 2022), and validated in the Mass General Brigham Biobank (MGBB) based on a previous published pipeline (Misra, A et al., Nat Commun, Feb 2025). For each condition, the PRSMIX+ score performs at least as well as the strongest individual input scores in terms of association, variance, and discrimination strength. The condition-specific PRSMix+ scores are converted into percentile and stratified into risk categories (e.g., high, above average, average, and below average). Relative effects are calculated for each risk category per disease, with relative effects by genetic similarity group are provided in the supporting materials of the report.
The initial sequencing component of this test was performed by the Broad Clinical Laboratory of the Broad Institute (27 Blue Sky Drive, Burlington, MA 01803; CLIA#22D2055652), and interpretive algorithms and clinical reports were generated by the Laboratory for Molecular Medicine at Mass General Brigham Personalized Medicine (LMM, 65 Landsdowne St, Cambridge, MA 02139; 617-768-8500; CLIA#22D1005307). This test has not been cleared or approved by the U.S. Food and Drug Administration (FDA). The FDA has determined that such clearance or approval is not required for clinical use.
