Efficient development of certified diagnostic laboratory developed tests using proteomic data
Background
Proteomic technology is a powerful biological tool with established methods for identifying proteomic biomarkers, but the development of certified diagnostic clinical tests based on proteomic biomarkers can be time-consuming, prone to overfitting issues, and difficult to navigate. We demonstrate the utility of combining pipeline tools, statistical learning techniques, and a knowledge base of in-silico proteomic datasets into a reproducible workflow that allows for efficient development of LDT-certifiable tests using SomaScan® technology.
Methods
Data pipeline and analysis tools were developed using R, in conjunction with proteomic measurements obtained using the SomaScan Platform. The tools take the analyst from data processing and QC through identification of optimized models for prediction of clinical endpoints, and then through validation on a hold-out test set. The tools include an assessment of model robustness against sample handling issues, longitudinal stability, the impacts of assay noise on model performance, effects of putative interferents, and risk of failure during CLIA validation in the lab. Real-life examples of clinical applications demonstrate the effectiveness of the tool in reducing analysis time and increasing model accuracy.
Results
Analysis time for identifying the optimal proteomic model to validation was reduced by at least 80%, with decreased prediction variability by up to 90%. In at least 75% of cases, application of in-silico data allows for tuning of predictive models to ensure robustness in a variety of everyday settings. This tool has led to 16 LDT certified SomaLogic tests in the last 3 years, ranging from anthropometric
measurements to cardiovascular- and cancer-risk predictions.
Conclusions
Not only are powerful, proteomics-driven, diagnostic tests realizable, but they can be LDT certified in an efficient, reproducible manner and made to be robust to real-life variability. Efficient analysis tools allow us to leverage proteomic technology in new ways, leading to tests that can be used for precision medicine applications.
Authors
Y. Hagar
L.E. Alexander
C. Scheidel
A. Zhang
J. Gogain
C. Paterson
R. Ostroff
M.A. Hinterberg
SomaLogic Operating Co., Inc., Boulder, CO, USA
More posters
PosterComparison of Proteomic CV Risk to Established ASCVD 10-Year Risk Decision Points
The ASCVD pooled cohort equation (PCE) is well-established for CV risk assessment. Decision points for determining treatment plans are low, intermediate and high risk over 10 years, however this approach over and underestimates risk in certain subgroups. The validated CV Risk SomaSignal® Test (SST) provides 4-year risk probability of MACE allowing for timely assessment of risk, but the shorter timescale makes comparison to 10-year PCE risk less intuitive.
PosterStatin signature: using proteomics to detect pharmacological fingerprints
Using a previously described metacohort (n=5,575) of patients with increased CV risk, we hypothesized that PCE would stratify patients differently than the CV Risk SST, and that CV Risk score scaled to 10 years would yield an improved net reclassification index (NRI).
PosterUsing a proteomics-based cardiovascular risk test to identify systemic changes in a clinical trial of nonalcoholic fatty liver disease
Improvement in hepaKc inflammaKon, NAFLD acKvity score and fibrosis were associated with improved proteomic CV risk scores regardless of treatment provided.