Anonymized plasma Olink NPX matrices PCNSL (LOC network)

Processed Olink Proximity Extension Assay (PEA) Normalized Protein eXpression (NPX) values from newly diagnosed primary central nervous system lymphoma (PCNSL) patients enrolled in the French LOC network. This deposit supports the manuscript Plasma proteomics for predictive and prognostic stratification in newly diagnosed primary central nervous system lymphoma (PCNSL). Analysis code (consensus clustering, immune-signature scoring, survival models, and transcriptomics integration) is available at: https://github.com/tamelenak/pcnsl-plasma-proteomics Clinical annotations (age, sex, Karnofsky performance status, overall and progression-free survival, treatment response) are not included in this file and remain available upon reasonable request subject to LOC network data-sharing regulations. Cohort Platform: Olink PEA Reveal panel (1,007 proteins after platform QC). Samples: 172 rows = 164 unique patients + 8 technical replicate measurements. Batches: training (discovery cohort, enrolled January 2016 – June 2019) and validation (held-out cohort, enrolled July 2019 – December 2022). Replicate rows have an empty batch field. Anonymization All original patient identifiers and cross-modality linkers (MRI, omics_ID) were removed. Samples were assigned new pseudonymous codes (PCNSL- + 8 hexadecimal characters). Row order was shuffled; codes are not relatable to the original integer IDs without the withheld crosswalk. No demographic or outcome variables are included. NPX values are unchanged from the processed Olink readout. CSV format Metadata columns (3) Column Type Description sample_id string Pseudonymous sample identifier batch string training, validation, or empty (replicates) replicate_of string For the 8 replicate rows: sample_id of the corresponding patient sample; empty otherwise Protein columns (1007) Gene-symbol column headers (e.g. CXCL9, IL6, IL10). Values are Olink NPX on the log₂ scale as returned by the Olink analysis pipeline. Immune signatures (defined in the manuscript / GitHub code) Four pre-specified plasma immune signatures are scored as the mean of constituent NPX values: IFN/Chemokine: CXCL9, CXCL10, IFNG, CCL19, CCL20, CCL3, CCL4 Checkpoint/Cytokine: IL10, IL6, PDCD1, PDCD1LG2, LAG3, CD86, TNF Innate/Inflammatory: IL1B, IL6, TNF, CCL20, CCL3, CCL4, FASLG Integrated Immune: union of the above (17 proteins, including TNFRSF13C) Reproducibility note The GitHub notebooks include survival and prediction analyses that require clinical outcome variables not present in this deposit. This file enables inspection of the plasma proteome, signature scoring on NPX, and unsupervised analyses that do not require outcomes. Full reproduction of prognostic and predictive figures requires the controlled clinical dataset (on request). Contact For access to clinical annotations or questions about data use, contact the corresponding author listed in the manuscript.