Glomerular diseases, which are currently diagnosed using an invasive renal biopsy, encompass numerous disease subtypes that often display similar clinical manifestations even though they have different therapeutic regimes. Therefore, a noninvasive assay is needed to classify and guide the treatment of glomerular diseases. Here, we develop and apply a high-throughput and quantitative microarray platform to characterize the immunoglobulin proteome in the serum from 419 healthy and diseased patients. The immunoglobulin proteome-clinical variable correlation network revealed novel pathological mechanisms of glomerular diseases. Furthermore, an immunoglobulin proteome-multivariate normal distribution (IP-MiND) mathematical model based on the correlation network classified healthy volunteers and patients with idiopathic membranous nephropathy with an average recall of 48% (23-80%) in the discovery cohort and 64% (63-65%) in an independent validation cohort. Our results demonstrate the translational utility of our microarray platform to glomerular diseases as well as its clinical potential in characterizing other human diseases.
Keywords: autoantibody; biomarker; disease classification; disease modeling; immunoglobulins; immunotherapy; protein microarrays; serum.