(1083-D) Implementation of Laboratory Automation to Support Scaling of Cancer Patients Testing using OmniSeq® INSIGHT Molecular Diagnostic Test

Abstract: *JA and HTD contributed equally to this work.

Background: OmniSeq is a CAP/CLIA and NYS CLEP certified reference laboratory that utilizes OmniSeq® INSIGHT to perform molecular diagnostic testing on cancer patients. OmniSeq® INSIGHT is a next-generation sequencing (NGS) test based on comprehensive genomic and immune profiling (CGIP) test for the detection of genomic variants, signatures, HLA Class I genotypes, and immune gene expression in formalin-fixed paraffin-embedded (FFPE) tumor tissue. For the NGS assay, the preparation of DNA and RNA sequencing libraries is a vital part in obtaining high-quality sequencing results. The library preparation workflow is complicated, laborious and involves multiple repetitive steps, making this workflow suitable for automated solutions thereby decreasing operator’s hands-on time, improving precision, eliminating time-consuming repetitive manual pipetting, and reducing the risk of human error. To address these issues, we have implemented an automated post-PCR workflow for the OmniSeq® INSIGHT test that uses the Hamilton® SBS STARTM MOA liquid handler to prepare DNA and RNA libraries.

Methods: To validate the automated post-PCR workflow, gold standard FFPE samples with a minimum of 20% tumor nuclei were selected for manual and automated TruSight® Oncology (TSO500) library preparation. A total of 76 unique RNA samples and 109 unique DNA samples representing 20 tumor types from different tissue sources were manually prepared through the pre-PCR and post-PCR stages. Hamilton® instruments (n=2) were then utilized to prepare post-PCR sequencing libraries from the common pre-PCR aliquots. The final normalized libraries were denatured and pooled for sequencing using NovaSeq®6000 sequencers with analysis performed using the TSO500 bioinformatics pipeline. Automated sequencing results were compared to the manual results from same samples for accuracy (positive percent agreement-PPA and negative percent agreement-NPA), and precision (average positive agreement-APA and average negative agreement-ANA) for all variant categories.

Results: In the accuracy study, the concordance of each variant category was: 1) small variants ranged from 97.7%-99.1% for PPA and 100.0% for NPA; 2) Copy number (gain/loss) ranged from 88.5%-100.0% for PPA and 99.5-99.8% for NPA. The lower 88.5% copy gain PPA was attributed to cases that had fold-change values bordering the defined copy number calling thresholds. However, the Pearson R correlation between automation and manual methods was 0.99 (R2=0.99); 3) Tumor Mutational Burden (TMB)-High demonstrated a 97.5% PPA and 96.9% NPA. In the within-run precision study, the concordance of each category ranges from 96.7%-100.0% for APA and 99.3-100.0% for ANA. In the between-run precision study, the concordance of each category ranges from 97.3%-100% for APA and 97.1%-100.0% for ANA.

Conclusions: Our validation study demonstrates that the automated NGS library preparation post-PCR workflow using Hamilton® SBS STARTM MOA liquid handler produces the same high quality NGS libraries as the validated manual procedure. Overall, automation can support the scaling of OmniSeq® INSIGHT testing for cancer patients by reducing hands-on time, improving efficiency, and decreasing operator fatigue and error.