Okello-et-al-2026-A plasma protein biomarker signature that differentiates acute rheumatic fever from related clinical presentations-2
Description
Multiomic profiling of Acute Rheumatic Fever (ARF) and control individuals. ARF is a systemic inflammatory condition triggered by Group A Streptococcus infection, with timely diagnosis critical to prevent Rheumatic Heart Disease. ARF pathogenesis is poorly understood and diagnosis is based on clinical criteria. Here, we compared ARF cases and well-defined controls from two Uganda cohorts. We identified a 5-protein signature that discriminates ARF patients from clinically-similar conditions (receiver operating characteristic-area under the curve (ROC-AUC)=1.0, n=18 definite ARF vs n=9 known alternate diagnosis; ROC-AUC=0.97, n=18 definite ARF vs n=13 unknown alternate diagnosis), which retained very good diagnostic value in a validation cohort (ROC-AUC=0.83 n=26 definite ARF vs n=13 unknown alternate diagnosis). Pathway analysis identified the epithelial-mesenchymal transition pathway as highly-associated with acute ARF, suggesting that tissue damage and repair is central to ARF pathogenesis. Our findings require further validation, yet highlight the potential for proteomics to identify clinically useful diagnostic biomarkers that would revolutionise ARF diagnosis and treatment.
Files
Steps to reproduce
Proteomics The SomaScan™ proteomic profiling platform was used to assay 55ul of plasma collected from venous heparinised blood. SomaScan Assay data were normalised using internal hybridisation controls to mitigate bias caused by differential readout conditions in the individual microarrays due to eluate transfer, hybridisation, wash, and scan. This was followed by median signal normalisation across pooled calibrator replicates within each run to mitigate within-run technical variation in the calibrator signal prior to use in scaling calculations. The set of ratios of the calibrator reference value to the median of calibrator replicates for each SOMAmer reagent was used to calculate plate scale and calibration scale.; scale factor acceptance criteria per plate were predefined as between 0.4-2.5 and calibration scale factors between 0.6-1.4. In total, data were obtained for 7596 SOMAmer reagents. We retained SOMAmers corresponding to human (p=7335) proteins (p=7288) that passed Somalogic quality control checks. The resulting 6664 SOMAmers reagents (corresponding to 5911 unique Uniprot protein identifiers) were used for analysis. Open-source microarray analysis software from the R/Bioconductor consortium was used to analyse the SomaScan™ microarray data (http://www.bioconductor.org/). Proteins with different relative plasma concentrations between grouped by disease stages were identified using a robust linear regression model and a moderated t statistic as implemented in the R package “limma”. Metabolomics Untargeted metabolomic profiling was performed on plasma samples using liquid chromatography coupled to high-resolution mass-spectrometry (LC-MS). All identified metabolites were annotated using appropriate orthogonal analytical techniques applied to the metabolite of interest against a chemical reference standard. Covariate-adjusted generalized linear model (GLM - with identity link function & normal distribution, and adjusting for collection site, age & sex) was applied to the log10 transformed data of each of the 768 metabolites. Two statistical comparisons were performed. Firstly, individuals with definite ARF diagnosis against HC, and then definitive ARF against either known or unknown alternate diagnosis. Flow Cytometry Flow cytometry data of whole blood samples were acquired on an LSR Fortessa and analysed manually using Flowjo software (v9.9) following a pre-defined gating strategy. In addition, immunophenotyping using unbiased automated gating was undertaken on the same samples using flowCut, flowDensity, and flowTypeFilter as well as biomarker visualisation (RchyOptimyx).
Institutions
- The University of Western AustraliaWestern Australia, Perth