Prediction of Schizophrenia Diagnosis by Integration of Genetically Correlated Conditions and Traits

Abstract

Schizophrenia is genetically heterogeneous and comorbid with many conditions. In this study, we explored polygenic scores (PGSs) from genetically related conditions and traits to predict schizophrenia diagnosis using both logistic regression and deep neural network (DNN) models. We used the combined Molecular Genetics of Schizophrenia and Swedish Schizophrenia Case Control Study (MGS+SSCCS) data for training and testing the models, and used the Clinical Antipsychotic Trials for Intervention Effectiveness (CATIE) data as independent validation. We screened 28 conditions and traits comorbid with schizophrenia to identify traits as potential predictors and used LASSO regression to select predictors for model construction. We investigated how PGS calculation influenced model performance. We found that the inclusion of comorbid traits improved model performance and PGSs calculated from two traits were more generalizable in independent validation. With a DNN model using 19 PGS predictors, we accomplished a prediction accuracy of 0.813 and an AUC of 0.905 in the MGS+SSCCS data. When this model was validated with the CATIE data, it achieved an accuracy of 0.721 and AUC of 0.747. Our results indicate that PGSs alone may not be sufficient to predict schizophrenia accurately and the inclusion of behavioral and clinical data may be necessary for more accurate prediction model.

Artificial image objects for classification of schizophrenia with GWAS-selected SNVs and convolutional neural network

“…The AUCs from these studies varied substantially, from 0.54 to 0.95. The AIO results reported here fared better than most the studies reviewed, including our previous study, 18 which used not only the PRSs of schizophrenia but also PRSs from other comorbid traits. These results indicated that genetic data, once transformed into AIOs, could be used effectively to classify the diagnosed patients and unaffected controls.…”

“…In the literature, there were multiple studies that used genetic markers to predict schizophrenia diagnosis, 17 including a study from our group. 18 Most of these studies used an SVM algorithm, and the variables were mostly individual SNVs. Only two studies, one of which was our previous study, used external validation samples.…”

Machine learning for genetic prediction of psychiatric disorders: a systematic review

“…One publication [3] used both manual and random elements for search, and is counted in both categories. Manual tuning by Chen et al (2019) is implied through reported values which were attempted for hyperparameters, but not explicitly stated [5]. Hyperparameters searched systematically using a given set of values are denoted as grid search.…”

Clinical Approaches to Late-Onset Psychosis

“…These results still need to be replicated by other cohorts or similar genetic studies, but they suggest different approaches to discover genetic backgrounds in early-onset psychosis and those in late-onset psychosis. There is a high amount of genetic research related to schizophrenia [ 61 , 62 , 63 ], but the underlying difference between schizophrenic individuals with the typical onset age and those with late-onset schizophrenia means that more investigation is needed in this field. Another approach to find the pathophysiology of late-onset psychosis targets the shared liability between late-onset psychosis and the neurodegenerative disorder accompanying psychotic symptoms.…”