Classification of Schizophrenia, Bipolar Disorder and Major Depressive Disorder With Comorbid Traits and Deep Learning Algorithms

Authors

Xiangning Chen
Yimei Lu
Joan Manuel Cue
Mira V Han
Vishwajit L Nimgaonkar
Daniel R Weinberger
Shizhong Han
Zhongming Zhao
Jingchun Chen

Publication Date

2-5-2025

Journal

Schizophrenia

Abstract

Many psychiatric disorders share genetic liabilities, but whether these shared liabilities can be utilized to classify and differentiate psychiatric disorders remains unclear. In this study, we use polygenic risk scores (PRSs) of 42 traits comorbid with schizophrenia (SCZ), bipolar disorder (BIP), and major depressive disorder (MDD) to evaluate their utilities. We found that combining target specific PRS with PRSs of comorbid traits can improve the classification of the target disorders. Importantly, without inclusion of PRSs from targeted disorders, we can still classify SCZ (accuracy 0.710 ± 0.008, AUC 0.789 ± 0.011), BIP (accuracy 0.782 ± 0.006, AUC 0.852 ± 0.004), and MDD (accuracy 0.753 ± 0.019, AUC 0.822 ± 0.010). Furthermore, PRSs from comorbid traits alone can effectively differentiate unaffected controls and patients with SCZ, BIP, and MDD (accuracy 0.861 ± 0.003, AUC 0.961 ± 0.041). Our results demonstrate that shared liabilities can be used effectively to improve the classification and differentiation of these disorders. The finding that PRSs from comorbid traits alone can classify and differentiate SCZ, BIP and MDD reasonably well implies that a majority of the risk variants composing target PRSs are shared with comorbid traits. Overall, our results suggest that a data-driven approach may be feasible to classify and differentiate these disorders.

Keywords

Psychiatric disorders, Genetics of the nervous system

DOI

10.1038/s41537-025-00564-7

PMID

39910091

PMCID

PMC11799204

PubMedCentral® Posted Date

2-5-2025

PubMedCentral® Full Text Version

Post-print

Published Open-Access

yes