Table 1_Integrative plasma proteomics and myeloid- interferon profiling reveal an AI-valid

16 plasma samples from an Indian cohort were profiled using LC-MS/MS proteomics and Olink cytokine assays to distinguish Systemic Lupus Erythematosus (SLE) flare from remission states. The dataset includes a 5-protein biosignature and was used to train a Random Forest classifier achieving a mean AUROC of 0.96. Data are available via ProteomeXchange with identifier PXD075349 and were last updated on 2026-06-03.

Use Cases

• Training machine learning classifiers for disease activity prediction based on the identified 5-protein biosignature (COL18A1, HYOU1, IGHG4, FLNA, SH3BGRL3).
• Validating vascular-endothelial and myeloid-interferon pathway dysregulation in autoimmune flare states based on differential protein expression.
• Benchmarking biomarker sensitivity and specificity using the reported COL18A1 cut-off with 100% specificity and 87.5% sensitivity.
• Conducting comparative proteomic studies in underrepresented populations based on data from a South Asian cohort.

Strengths

• Includes a machine learning model validated with leave-one-out cross-validation, permutation testing, and bootstrapped AUROC confidence intervals.
• Reports a specific biomarker cut-off (COL18A1) with 100% specificity and 87.5% sensitivity for disease activity.
• Data is publicly available via the ProteomeXchange repository with identifier PXD075349 under a CC-BY-4.0 license.

Limitations

• The cohort size is 16 participants, which is a small sample for a discovery-phase study.
• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset is 5.1 MB in size, suggesting a limited scale of underlying data.

Provenance

Source

figshare

Collection Method

Plasma proteomic profiling via LC-MS/MS integrated with targeted cytokine quantification using Olink Proximity Extension Assay (PEA).

Freshness

Last updated 2026-06-03 06:01:31; freshness should be verified.

Geography

Indian cohort