Fig. 5: Association between protein abundance and WOMAC knee pain subscore in non-IPS regressed data using linear regression modelling. | Nature Communications

Fig. 5: Association between protein abundance and WOMAC knee pain subscore in non-IPS regressed data using linear regression modelling.

From: Large-scale molecular endotype discovery in synovial fluid reveals osteoarthritis as a single biological continuum

Fig. 5: Association between protein abundance and WOMAC knee pain subscore in non-IPS regressed data using linear regression modelling.The alternative text for this image may have been generated using AI.

Protein abundance was measured in 805 OA samples where WOMAC knee pain subscore data was available (Combined: 748 spun, 57 unspun), adjusted for spin-status (using ComBat) and then age and biological sex. A Volcano plot showing beta estimates per standard deviation change in protein expression for proteins associated with WOMAC knee pain subscore in the Combined dataset, with Benjamini-Hochberg adjusted p-values (padj). Proteins in red are positively associated, those in blue are negatively associated, with increasing WOMAC knee pain subscore at padj ≤ 0.05. Top 30 associated proteins, for each direction, ordered by padj are labelled. In green are two proteins that were significant in Replication and Combined datasets (at padj ≤ 0.05), including after cohort adjustment, see Supplementary Fig. 7. B A scatter plot of beta estimates from linear regression models of the associations between protein abundance and WOMAC knee pain in non-IPS analyses is shown for Discovery and Replication datasets with significantly associated proteins (at padj ≤ 0.05) in different groups shown in different colours (see key). Pearson correlation coefficient and p-value (unadjusted) are presented for the correlation between beta estimates generated in non-IPS regressed analyses using Discovery and Replication datasets. C Scatter plots of WOMAC pain subscore against NOE2 or NAR3 protein abundance (transformed by natural logarithms) in OA participants using Combined, spin-status corrected, non-IPS regressed data. Beta estimates and p-values (unadjusted) are presented for linear models adjusted for age and biological sex. D Bubble plot of significantly enriched pathways (padj <0.05) using the Hallmark Gene set for proteins associated with WOMAC knee pain subscore by Replication and Combined datatsets not adjusted for IPS or cohort. No pathways were significantly enriched at padj <0.05 in the Discovery dataset. OA osteoarthritis, SD standard deviation, IPS intracellular protein score, WOMAC Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC, 0 = no pain, 100 = worst possible pain), noelin-2 (NOE2); ecto-ADP-ribosyltransferase 3 (NAR3), (padj) adjusted p-value, NES normalised enrichment score. Full list of proteins available in Source Data file 8.

Back to article page