Predictive modeling of schizophrenia from genomic data: Comparison of polygenic risk score with kernel support vector machines approach

Discrimination between healthy participants and people with panic disorder based on polygenic scores for psychiatric disorders and for intermediate phenotypes using machine learning

OBJECTIVE

Panic disorder is a modestly heritable condition. Currently, diagnosis is based only on clinical symptoms; identifying objective biomarkers and a more reliable diagnostic procedure is desirable. We investigated whether people with panic disorder can be reliably diagnosed utilizing combinations of multiple polygenic scores for psychiatric disorders and their intermediate phenotypes, compared with single polygenic score approaches, by applying specific machine learning techniques.

METHODS

Polygenic scores for 48 psychiatric disorders and intermediate phenotypes based on large-scale genome-wide association studies (n = 7556-1,131,881) were calculated for people with panic disorder (n = 718) and healthy controls (n = 1717). Discrimination between people with panic disorder and healthy controls was based on the 48 polygenic scores using five methods for classification: logistic regression, neural networks, quadratic discriminant analysis, random forests and a support vector machine. Differences in discrimination accuracy (area under the curve) due to an increased number of polygenic score combinations and differences in the accuracy across five classifiers were investigated.

RESULTS

All five classifiers performed relatively well for distinguishing people with panic disorder from healthy controls by increasing the number of polygenic scores. Of the 48 polygenic scores, the polygenic score for anxiety UK Biobank was the most useful for discrimination by the classifiers. In combinations of two or three polygenic scores, the polygenic score for anxiety UK Biobank was included as one of polygenic scores in all classifiers. When all 48 polygenic scores were used in combination, the greatest areas under the curve significantly differed among the five classifiers. Support vector machine and logistic regression had higher accuracy than quadratic discriminant analysis and random forests. For each classifier, the greatest area under the curve was 0.600 ± 0.030 for logistic regression (polygenic score combinations N = 14), 0.591 ± 0.039 for neural networks (N = 9), 0.603 ± 0.033 for quadratic discriminant analysis (N = 10), 0.572 ± 0.039 for random forests (N = 25) and 0.617 ± 0.041 for support vector machine (N = 11). The greatest areas under the curve at the best polygenic score combination significantly differed among the five classifiers. Random forests had the lowest accuracy among classifiers. Support vector machine had higher accuracy than neural networks.

CONCLUSIONS

These findings suggest that increasing the number of polygenic score combinations up to approximately 10 effectively improved the discrimination accuracy and that support vector machine exhibited greater accuracy among classifiers. However, the discrimination accuracy for panic disorder, when based solely on polygenic score combinations, was found to be modest.

A Systematic Review of Genetics- and Molecular-Pathway-Based Machine Learning Models for Neurological Disorder Diagnosis

The process of identification and management of neurological disorder conditions faces challenges, prompting the investigation of novel methods in order to improve diagnostic accuracy. In this study, we conducted a systematic literature review to identify the significance of genetics- and molecular-pathway-based machine learning (ML) models in treating neurological disorder conditions. According to the study's objectives, search strategies were developed to extract the research studies using digital libraries. We followed rigorous study selection criteria. A total of 24 studies met the inclusion criteria and were included in the review. We classified the studies based on neurological disorders. The included studies highlighted multiple methodologies and exceptional results in treating neurological disorders. The study findings underscore the potential of the existing models, presenting personalized interventions based on the individual's conditions. The findings offer better-performing approaches that handle genetics and molecular data to generate effective outcomes. Moreover, we discuss the future research directions and challenges, emphasizing the demand for generalizing existing models in real-world clinical settings. This study contributes to advancing knowledge in the field of diagnosis and management of neurological disorders.

Immunogenetics of Systemic Sclerosis

Systemic sclerosis (SSc) is a rare autoimmune connective tissue disorder characterized by massive fibrosis, vascular damage, and immune imbalance. Advances in rheumatology and immunology over the past two decades have led to a redefinition of systemic sclerosis, shifting from its initial perception as primarily a "hyperfibrotic" state towards a recognition of systemic sclerosis as an immune-mediated disease. Consequently, the search for genetic markers has transitioned from focusing on fibrotic mechanisms to exploring immune regulatory pathways. Immunogenetics, an emerging field at the intersection of immunology, molecular biology, and genetics has provided valuable insights into inherited factors that influence immunity. Data from genetic studies conducted thus far indicate that alterations in genetic messages can significantly impact disease risk and progression. While certain genetic variations may confer protective effects, others may exacerbate disease susceptibility. This paper presents a comprehensive review of the most relevant genetic changes that influence both the risk and course of systemic sclerosis. Special emphasis is placed on factors regulating the immune response, recognizing their pivotal role in the pathogenesis of the disease.

Replication of a neuroimaging biomarker for striatal dysfunction in psychosis

To bring biomarkers closer to clinical application, they should be generalizable, reliable, and maintain performance within the constraints of routine clinical conditions. The functional striatal abnormalities (FSA), is among the most advanced neuroimaging biomarkers in schizophrenia, trained to discriminate diagnosis, with post-hoc analyses indicating prognostic properties. Here, we attempt to replicate its diagnostic capabilities measured by the area under the curve (AUC) in receiver operator characteristic curves discriminating individuals with psychosis (n = 101) from healthy controls (n = 51) in the Human Connectome Project for Early Psychosis. We also measured the test-retest (run 1 vs 2) and phase encoding direction (i.e., AP vs PA) reliability with intraclass correlation coefficients (ICC). Additionally, we measured effects of scan length on classification accuracy (i.e., AUCs) and reliability (i.e., ICCs). Finally, we tested the prognostic capability of the FSA by the correlation between baseline scores and symptom improvement over 12 weeks of antipsychotic treatment in a separate cohort (n = 97). Similar analyses were conducted for the Yeo networks intrinsic connectivity as a reference. The FSA had good/excellent diagnostic discrimination (AUC = 75.4%, 95% CI = 67.0-83.3%; in non-affective psychosis AUC = 80.5%, 95% CI = 72.1-88.0%, and in affective psychosis AUC = 58.7%, 95% CI = 44.2-72.0%). Test-retest reliability ranged between ICC = 0.48 (95% CI = 0.35-0.59) and ICC = 0.22 (95% CI = 0.06-0.36), which was comparable to that of networks intrinsic connectivity. Phase encoding direction reliability for the FSA was ICC = 0.51 (95% CI = 0.42-0.59), generally lower than for networks intrinsic connectivity. By increasing scan length from 2 to 10 min, diagnostic classification of the FSA increased from AUC = 71.7% (95% CI = 63.1-80.3%) to 75.4% (95% CI = 67.0-83.3%) and phase encoding direction reliability from ICC = 0.29 (95% CI = 0.14-0.43) to ICC = 0.51 (95% CI = 0.42-0.59). FSA scores did not correlate with symptom improvement. These results reassure that the FSA is a generalizable diagnostic - but not prognostic - biomarker. Given the replicable results of the FSA as a diagnostic biomarker trained on case-control datasets, next the development of prognostic biomarkers should be on treatment-response data.

Genomic Machine Learning Meta-regression: Insights on Associations of Study Features With Reported Model Performance

Many studies have been conducted with the goal of correctly predicting diagnostic status of a disorder using the combination of genomic data and machine learning. It is often hard to judge which components of a study led to better results and whether better reported results represent a true improvement or an uncorrected bias inflating performance. We extracted information about the methods used and other differentiating features in genomic machine learning models. We used these features in linear regressions predicting model performance. We tested for univariate and multivariate associations as well as interactions between features. Of the models reviewed, 46% used feature selection methods that can lead to data leakage. Across our models, the number of hyperparameter optimizations reported, data leakage due to feature selection, model type, and modeling an autoimmune disorder were significantly associated with an increase in reported model performance. We found a significant, negative interaction between data leakage and training size. Our results suggest that methods susceptible to data leakage are prevalent among genomic machine learning research, resulting in inflated reported performance. Best practice guidelines that promote the avoidance and recognition of data leakage may help the field avoid biased results.

Replication of a neuroimaging biomarker for striatal dysfunction in psychosis

To bring biomarkers closer to clinical application, they should be generalizable, reliable, and maintain performance within the constraints of routine clinical conditions. The functional striatal abnormalities (FSA), is among the most advanced neuroimaging biomarkers in schizophrenia, trained to discriminate diagnosis, with post-hoc analyses indicating prognostic properties. Here, we attempt to replicate its diagnostic capabilities measured by the area under the curve (AUC) in receiver operator characteristic curves discriminating individuals with psychosis (n=101) from healthy controls (n=51) in the Human Connectome Project for Early Psychosis. We also measured the test-retest (run 1 vs 2) and phase encoding direction (i.e., AP vs PA) reliability with intraclass correlation coefficients (ICC). Additionally, we measured effects of scan length on classification accuracy (i.e., AUCs) and reliability (i.e., ICCs). Finally, we tested the prognostic capability of the FSA by the correlation between baseline scores and symptom improvement over 12 weeks of antipsychotic treatment in a separate cohort (n=97). Similar analyses were conducted for the Yeo networks intrinsic connectivity as a reference. The FSA had good/excellent diagnostic discrimination (AUC=75.4%, 95%CI=67.0%-83.3%; in non-affective psychosis AUC=80.5%, 95%CI=72.1-88.0%, and in affective psychosis AUC=58.7%, 95%CI=44.2-72.0%). Test-retest reliability ranged between ICC=0.48 (95%CI=0.35-0.59) and ICC=0.22 (95%CI=0.06-0.36), which was comparable to that of networks intrinsic connectivity. Phase encoding direction reliability for the FSA was ICC=0.51 (95%CI=0.42-0.59), generally lower than for networks intrinsic connectivity. By increasing scan length from 2 to 10 minutes, diagnostic classification of the FSA increased from AUC=71.7% (95%CI=63.1%-80.3%) to 75.4% (95%CI=67.0%-83.3%) and phase encoding direction reliability from ICC=0.29 (95%CI=0.14-0.43) to ICC=0.51 (95%CI=0.42-0.59). FSA scores did not correlate with symptom improvement. These results reassure that the FSA is a generalizable diagnostic - but not prognostic - biomarker. Given the replicable results of the FSA as a diagnostic biomarker trained on case-control datasets, next the development of prognostic biomarkers should be on treatment-response data.

Replication of a neuroimaging biomarker for striatal dysfunction in psychosis

To bring biomarkers closer to clinical application, they should be generalizable, reliable, and maintain performance within the constraints of routine clinical conditions. The functional striatal abnormalities (FSA), is among the most advanced neuroimaging biomarkers in schizophrenia, trained to discriminate diagnosis, with post-hoc analyses indicating prognostic properties. Here, we attempt to replicate its diagnostic capabilities measured by the area under the curve (AUC) in receiver operator characteristic curves discriminating individuals with psychosis (n=101) from healthy controls (n=51) in the Human Connectome Project for Early Psychosis. We also measured the test-retest (run 1 vs 2) and phase encoding direction (i.e., AP vs PA) reliability with intraclass correlation coefficients (ICC). Additionally, we measured effects of scan length on classification accuracy (i.e., AUCs) and reliability (i.e., ICCs). Finally, we tested the prognostic capability of the FSA by the correlation between baseline scores and symptom improvement over 12 weeks of antipsychotic treatment in a separate cohort (n=97). Similar analyses were conducted for the Yeo networks intrinsic connectivity as a reference. The FSA had good/excellent diagnostic discrimination (AUC=75.4%, 95%CI=67.0%-83.3%; in non-affective psychosis AUC=80.5%, 95%CI=72.1-88.0%, and in affective psychosis AUC=58.7%, 95%CI=44.2-72.0%). Test-retest reliability ranged between ICC=0.48 (95%CI=0.35-0.59) and ICC=0.22 (95%CI=0.06-0.36), which was comparable to that of networks intrinsic connectivity. Phase encoding direction reliability for the FSA was ICC=0.51 (95%CI=0.42-0.59), generally lower than for networks intrinsic connectivity. By increasing scan length from 2 to 10 minutes, diagnostic classification of the FSA increased from AUC=71.7% (95%CI=63.1%-80.3%) to 75.4% (95%CI=67.0%-83.3%) and phase encoding direction reliability from ICC=0.29 (95%CI=0.14-0.43) to ICC=0.51 (95%CI=0.42-0.59). FSA scores did not correlate with symptom improvement. These results reassure that the FSA is a generalizable diagnostic - but not prognostic - biomarker. Given the replicable results of the FSA as a diagnostic biomarker trained on case-control datasets, next the development of prognostic biomarkers should be on treatment-response data.

Learning high-order interactions for polygenic risk prediction

Within the framework of precision medicine, the stratification of individual genetic susceptibility based on inherited DNA variation has paramount relevance. However, one of the most relevant pitfalls of traditional Polygenic Risk Scores (PRS) approaches is their inability to model complex high-order non-linear SNP-SNP interactions and their effect on the phenotype (e.g. epistasis). Indeed, they incur in a computational challenge as the number of possible interactions grows exponentially with the number of SNPs considered, affecting the statistical reliability of the model parameters as well. In this work, we address this issue by proposing a novel PRS approach, called High-order Interactions-aware Polygenic Risk Score (hiPRS), that incorporates high-order interactions in modeling polygenic risk. The latter combines an interaction search routine based on frequent itemsets mining and a novel interaction selection algorithm based on Mutual Information, to construct a simple and interpretable weighted model of user-specified dimensionality that can predict a given binary phenotype. Compared to traditional PRSs methods, hiPRS does not rely on GWAS summary statistics nor any external information. Moreover, hiPRS differs from Machine Learning-based approaches that can include complex interactions in that it provides a readable and interpretable model and it is able to control overfitting, even on small samples. In the present work we demonstrate through a comprehensive simulation study the superior performance of hiPRS w.r.t. state of the art methods, both in terms of scoring performance and interpretability of the resulting model. We also test hiPRS against small sample size, class imbalance and the presence of noise, showcasing its robustness to extreme experimental settings. Finally, we apply hiPRS to a case study on real data from DACHS cohort, defining an interaction-aware scoring model to predict mortality of stage II-III Colon-Rectal Cancer patients treated with oxaliplatin.

Machine Learning and Non-Affective Psychosis: Identification, Differential Diagnosis, and Treatment

PURPOSE OF REVIEW

This review will cover the most relevant findings on the use of machine learning (ML) techniques in the field of non-affective psychosis, by summarizing the studies published in the last three years focusing on illness detection and treatment.

RECENT FINDINGS

Multiple ML tools that include mostly supervised approaches such as support vector machine, gradient boosting, and random forest showed promising results by applying these algorithms to various sources of data: socio-demographic information, EEG, language, digital content, blood biomarkers, neuroimaging, and electronic health records. However, the overall performance, in the binary classification case, varied from 0.49, which is to be considered very low (i.e., noise), to over 0.90. These results are fully justified by different factors, some of which may be attributable to the preprocessing of the data, the wide variety of the data, and the a-priori setting of hyperparameters. One of the main limitations of the field is the lack of stratification of results based on biological sex, given that psychosis presents differently in men and women; hence, the necessity to tailor identification tools and data analytic strategies. Timely identification and appropriate treatment are key factors in reducing the consequences of psychotic disorders. In recent years, the emergence of new analytical tools based on artificial intelligence such as supervised ML approaches showed promises as a potential breakthrough in this field. However, ML applications in everyday practice are still in its infancy.

Non-linear machine learning models incorporating SNPs and PRS improve polygenic prediction in diverse human populations

Polygenic risk scores (PRS) are commonly used to quantify the inherited susceptibility for a trait, yet they fail to account for non-linear and interaction effects between single nucleotide polymorphisms (SNPs). We address this via a machine learning approach, validated in nine complex phenotypes in a multi-ancestry population. We use an ensemble method of SNP selection followed by gradient boosted trees (XGBoost) to allow for non-linearities and interaction effects. We compare our results to the standard, linear PRS model developed using PRSice, LDpred2, and lassosum2. Combining a PRS as a feature in an XGBoost model results in a relative increase in the percentage variance explained compared to the standard linear PRS model by 22% for height, 27% for HDL cholesterol, 43% for body mass index, 50% for sleep duration, 58% for systolic blood pressure, 64% for total cholesterol, 66% for triglycerides, 77% for LDL cholesterol, and 100% for diastolic blood pressure. Multi-ancestry trained models perform similarly to specific racial/ethnic group trained models and are consistently superior to the standard linear PRS models. This work demonstrates an effective method to account for non-linearities and interaction effects in genetics-based prediction models.

Open problems in human trait genetics

Genetic studies of human traits have revolutionized our understanding of the variation between individuals, and yet, the genetics of most traits is still poorly understood. In this review, we highlight the major open problems that need to be solved, and by discussing these challenges provide a primer to the field. We cover general issues such as population structure, epistasis and gene-environment interactions, data-related issues such as ancestry diversity and rare genetic variants, and specific challenges related to heritability estimates, genetic association studies, and polygenic risk scores. We emphasize the interconnectedness of these problems and suggest promising avenues to address them.

Performance Evaluation of Machine Learning Algorithm for Classification of Unintended Pregnancy among Married Women in Bangladesh

Intended pregnancy is one of the significant indicators of women's well-being. Globally, 74 million women become pregnant every year without planning. Unintended pregnancies account for 28% of all pregnancies among married women in Bangladesh. This study aimed to investigate the performance of six different machine learning (ML) algorithms applied to predict unintended pregnancies among married women in Bangladesh. From BDHS 2017-18, only 1129 pregnant women aged 15-49 were eligible for this study. An independent χ 2 test had performed before we considered six popular ML algorithms, such as logistic regression (LR), random forest (RF), support vector machine (SVM), k-nearest neighbor (KNN), naïve Bayes (NB), and elastic net regression (ENR) to predict the unintended pregnancy. Accuracy, sensitivity, specificity, Cohen's Kappa statistic, and area under curve (AUC) value were used as model evaluation. The bivariate analysis result showed that women aged 30-49 years, poor, not educated, and living in male-headed households had a higher percentage of unintended pregnancy. We found various performance parameters for the classification of unintended pregnancy: LR accuracy = 79.29%, LR AUC = 72.12%; RF accuracy = 77.81%, RF AUC = 72.17%; SVM accuracy = 76.92%, SVM AUC = 70.90%; KNN accuracy = 77.22%, KNN AUC = 70.27%; NB accuracy = 78%, NB AUC = 73.06%; and ENR accuracy = 77.51%, ENR AUC = 74.67%. Based on the AUC value, we can conclude that of all the ML algorithms we investigated, the ENR algorithm provides the most accurate classification for predicting unwanted pregnancy among Bangladeshi women. Our findings contribute to a better understanding of how to categorize pregnancy intentions among Bangladeshi women. As a result, the government can initiate an effective campaign to raise contraception awareness.

Prediction of transition to psychosis from an at-risk mental state using structural neuroimaging, genetic, and environmental data

INTRODUCTION

Psychosis is usually preceded by a prodromal phase in which patients are clinically identified as being at in an "At Risk Mental State" (ARMS). A few studies have demonstrated the feasibility of predicting psychosis transition from an ARMS using structural magnetic resonance imaging (sMRI) data and machine learning (ML) methods. However, the reliability of these findings is unclear due to possible sampling bias. Moreover, the value of genetic and environmental data in predicting transition to psychosis from an ARMS is yet to be explored.

METHODS

In this study we aimed to predict transition to psychosis from an ARMS using a combination of ML, sMRI, genome-wide genotypes, and environmental risk factors as predictors, in a sample drawn from a pool of 246 ARMS subjects (60 of whom later transitioned to psychosis). First, the modality-specific values in predicting transition to psychosis were evaluated using several: (a) feature types; (b) feature manipulation strategies; (c) ML algorithms; (d) cross-validation strategies, as well as sample balancing and bootstrapping. Subsequently, the modalities whose at least 60% of the classification models showed an balanced accuracy (BAC) statistically better than chance level were included in a multimodal classification model.

RESULTS AND DISCUSSION

Results showed that none of the modalities alone, i.e., neuroimaging, genetic or environmental data, could predict psychosis from an ARMS statistically better than chance and, as such, no multimodal classification model was trained/tested. These results suggest that the value of structural MRI data and genome-wide genotypes in predicting psychosis from an ARMS, which has been fostered by previous evidence, should be reconsidered.

Multi-modal deep learning of functional and structural neuroimaging and genomic data to predict mental illness

Neuropsychiatric disorders such as schizophrenia are very heterogeneous in nature and typically diagnosed using self-reported symptoms. This makes it difficult to pose a confident prediction on the cases and does not provide insight into the underlying neural and biological mechanisms of these disorders. Combining neuroimaging and genomic data with a multi-modal 'predictome' paves the way for biologically informed markers and may improve prediction reliability. With that, we develop a multi-modal deep learning framework by fusing data from different modalities to capture the interaction between the latent features and evaluate their complementary information in characterizing schizophrenia. Our deep model uses structural MRI, functional MRI, and genome-wide polymorphism data to perform the classification task. It includes a multi-layer feed-forward network, an encoder, and a long short-term memory (LSTM) unit with attention to learn the latent features and adopt a joint training scheme capturing synergies between the modalities. The hybrid network also uses different regularizers for addressing the inherent overfitting and modality-specific bias in the multi-modal setup. Next, we run the network through a saliency model to analyze the learned features. Integrating modalities enhances the performance of the classifier, and our framework acquired 88% (P < 0.0001) accuracy on a dataset of 437 subjects. The trimodal accuracy is comparable to the state-of-the-art performance on a data collection of this size and outperforms the unimodal and bimodal baselines we compared. Model introspection was used to expose the salient neural features and genes/biological pathways associated with schizophrenia. To our best knowledge, this is the first approach that fuses genomic information with structural and functional MRI biomarkers for predicting schizophrenia. We believe this type of modality blending can better explain the disorder's dynamics by adding cross-modal prospects.Clinical Relevance- This study combinedly learns imaging and genomic features for the classification of schizophrenia. The data fusion scheme extracts modality interactions, and the saliency experiments report multiple functional and structural networks closely connected to the disorder.

A machine learning case-control classifier for schizophrenia based on DNA methylation in blood

Epigenetic dysregulation is thought to contribute to the etiology of schizophrenia (SZ), but the cell type-specificity of DNA methylation makes population-based epigenetic studies of SZ challenging. To train an SZ case-control classifier based on DNA methylation in blood, therefore, we focused on human genomic regions of systemic interindividual epigenetic variation (CoRSIVs), a subset of which are represented on the Illumina Human Methylation 450K (HM450) array. HM450 DNA methylation data on whole blood of 414 SZ cases and 433 non-psychiatric controls were used as training data for a classification algorithm with built-in feature selection, sparse partial least squares discriminate analysis (SPLS-DA); application of SPLS-DA to HM450 data has not been previously reported. Using the first two SPLS-DA dimensions we calculated a "risk distance" to identify individuals with the highest probability of SZ. The model was then evaluated on an independent HM450 data set on 353 SZ cases and 322 non-psychiatric controls. Our CoRSIV-based model classified 303 individuals as cases with a positive predictive value (PPV) of 80%, far surpassing the performance of a model based on polygenic risk score (PRS). Importantly, risk distance (based on CoRSIV methylation) was not associated with medication use, arguing against reverse causality. Risk distance and PRS were positively correlated (Pearson r = 0.28, P = 1.28 × 10^-12), and mediational analysis suggested that genetic effects on SZ are partially mediated by altered methylation at CoRSIVs. Our results indicate two innate dimensions of SZ risk: one based on genetic, and the other on systemic epigenetic variants.

The promise of machine learning in predicting treatment outcomes in psychiatry

For many years, psychiatrists have tried to understand factors involved in response to medications or psychotherapies, in order to personalize their treatment choices. There is now a broad and growing interest in the idea that we can develop models to personalize treatment decisions using new statistical approaches from the field of machine learning and applying them to larger volumes of data. In this pursuit, there has been a paradigm shift away from experimental studies to confirm or refute specific hypotheses towards a focus on the overall explanatory power of a predictive model when tested on new, unseen datasets. In this paper, we review key studies using machine learning to predict treatment outcomes in psychiatry, ranging from medications and psychotherapies to digital interventions and neurobiological treatments. Next, we focus on some new sources of data that are being used for the development of predictive models based on machine learning, such as electronic health records, smartphone and social media data, and on the potential utility of data from genetics, electrophysiology, neuroimaging and cognitive testing. Finally, we discuss how far the field has come towards implementing prediction tools in real-world clinical practice. Relatively few retrospective studies to-date include appropriate external validation procedures, and there are even fewer prospective studies testing the clinical feasibility and effectiveness of predictive models. Applications of machine learning in psychiatry face some of the same ethical challenges posed by these techniques in other areas of medicine or computer science, which we discuss here. In short, machine learning is a nascent but important approach to improve the effectiveness of mental health care, and several prospective clinical studies suggest that it may be working already.

A fuzzy gaussian rank aggregation ensemble feature selection method for microarray data

In Microarray Data, it is complicated to achieve more classification accuracy due to the presence of high dimensions, irrelevant and noisy data. And also It had more gene expression data and fewer samples. To increase the classification accuracy and the processing speed of the model, an optimal number of features need to extract, this can be achieved by applying the feature selection method. In this paper, we propose a hybrid ensemble feature selection method. The proposed method has two phases, filter and wrapper phase in filter phase ensemble technique is used for aggregating the feature ranks of the Relief, minimum redundancy Maximum Relevance (mRMR), and Feature Correlation (FC) filter feature selection methods. This paper uses the Fuzzy Gaussian membership function ordering for aggregating the ranks. In wrapper phase, Improved Binary Particle Swarm Optimization (IBPSO) is used for selecting the optimal features, and the RBF Kernel-based Support Vector Machine (SVM) classifier is used as an evaluator. The performance of the proposed model are compared with state of art feature selection methods using five benchmark datasets. For evaluation various performance metrics such as Accuracy, Recall, Precision, and F1-Score are used. Furthermore, the experimental results show that the performance of the proposed method outperforms the other feature selection methods.

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

Machine learning methods have been employed to make predictions in psychiatry from genotypes, with the potential to bring improved prediction of outcomes in psychiatric genetics; however, their current performance is unclear. We aim to systematically review machine learning methods for predicting psychiatric disorders from genetics alone and evaluate their discrimination, bias and implementation. Medline, PsycInfo, Web of Science and Scopus were searched for terms relating to genetics, psychiatric disorders and machine learning, including neural networks, random forests, support vector machines and boosting, on 10 September 2019. Following PRISMA guidelines, articles were screened for inclusion independently by two authors, extracted, and assessed for risk of bias. Overall, 63 full texts were assessed from a pool of 652 abstracts. Data were extracted for 77 models of schizophrenia, bipolar, autism or anorexia across 13 studies. Performance of machine learning methods was highly varied (0.48-0.95 AUC) and differed between schizophrenia (0.54-0.95 AUC), bipolar (0.48-0.65 AUC), autism (0.52-0.81 AUC) and anorexia (0.62-0.69 AUC). This is likely due to the high risk of bias identified in the study designs and analysis for reported results. Choices for predictor selection, hyperparameter search and validation methodology, and viewing of the test set during training were common causes of high risk of bias in analysis. Key steps in model development and validation were frequently not performed or unreported. Comparison of discrimination across studies was constrained by heterogeneity of predictors, outcome and measurement, in addition to sample overlap within and across studies. Given widespread high risk of bias and the small number of studies identified, it is important to ensure established analysis methods are adopted. We emphasise best practices in methodology and reporting for improving future studies.

Polygenic risk scores for genetic counseling in psychiatry: Lessons learned from other fields of medicine

Polygenic risk scores (PRS) may aid in the identification of individuals at-risk for psychiatric disorders, treatment optimization, and increase in prognostic accuracy. PRS may also add significant value to genetic counseling. Thus far, integration of PRSs in genetic counseling sessions remains problematic because of uncertainties in risk prediction and other concerns. Here, we review the current utility of PRSs in the context of clinical psychiatry. By comprehensively appraising the literature in other fields of medicine including breast cancer, Alzheimer's Disease, and cardiovascular disease, we outline several lessons learned that could be applied to future studies and may thus benefit the incorporation of PRS in psychiatric genetic counseling. These include integrating PRS with environmental factors (e.g. lifestyle), setting up large-scale studies, and applying reproducible methods allowing for cross-validation between cohorts. We conclude that psychiatry may benefit from experiences in these fields. PRS may in future have a role in genetic counseling in clinical psychiatric practice, by advancing prevention strategies and treatment decision-making, thus promoting quality of life for (potentially) affected individuals.

Challenges and Future Prospects of Precision Medicine in Psychiatry

Precision medicine is increasingly recognized as a promising approach to improve disease treatment, taking into consideration the individual clinical and biological characteristics shared by specific subgroups of patients. In specific fields such as oncology and hematology, precision medicine has already started to be implemented in the clinical setting and molecular testing is routinely used to select treatments with higher efficacy and reduced adverse effects. The application of precision medicine in psychiatry is still in its early phases. However, there are already examples of predictive models based on clinical data or combinations of clinical, neuroimaging and biological data. While the power of single clinical predictors would remain inadequate if analyzed only with traditional statistical approaches, these predictors are now increasingly used to impute machine learning models that can have adequate accuracy even in the presence of relatively small sample size. These models have started to be applied to disentangle relevant clinical questions that could lead to a more effective management of psychiatric disorders, such as prediction of response to the mood stabilizer lithium, resistance to antidepressants in major depressive disorder or stratification of the risk and outcome prediction in schizophrenia. In this narrative review, we summarized the most important findings in precision medicine in psychiatry based on studies that constructed machine learning models using clinical, neuroimaging and/or biological data. Limitations and barriers to the implementation of precision psychiatry in the clinical setting, as well as possible solutions and future perspectives, will be presented.

Application of deep canonically correlated sparse autoencoder for the classification of schizophrenia

BACKGROUND AND OBJECTIVE

Imaging genetics has been widely used to help diagnose and treat mental illness, e.g., schizophrenia, by combining magnetic resonance imaging of the brain and genomic information for comprehensive and systematic analysis. As a result, utilizing the correlation between magnetic resonance imaging of the brain and genomic information is becoming an important challenge.

METHODS

In this paper, the joint analysis of single nucleotide polymorphisms and functional magnetic resonance imaging is conducted for comprehensive study of schizophrenia. We developed a deep canonically correlated sparse autoencoder to classify schizophrenia patients from healthy controls, which can address the limitation of many existing methods such as canonical correlation analysis, deep canonical correlation analysis and sparse autoencoder.

RESULTS

The proposed deep canonically correlated sparse autoencoder can not only use complex nonlinear transformation and dimension reduction, but also achieve more accurate classifications. Our experiments showed the proposed method achieved an accuracy of 95.65% for SNP data sets and an accuracy of 80.53% for fMRI data sets.

CONCLUSIONS

Experiments demonstrated higher accuracy of using the proposed method over other conventional models when classifying schizophrenia patients and healthy controls.

Treatment-Resistant Schizophrenia: Insights From Genetic Studies and Machine Learning Approaches

Schizophrenia (SCZ) is a severe psychiatric disorder affecting approximately 23 million people worldwide. It is considered the eighth leading cause of disability according to the World Health Organization and is associated with a significant reduction in life expectancy. Antipsychotics represent the first-choice treatment in SCZ, but approximately 30% of patients fail to respond to acute treatment. These patients are generally defined as treatment-resistant and are eligible for clozapine treatment. Treatment-resistant patients show a more severe course of the disease, but it has been suggested that treatment-resistant schizophrenia (TRS) may constitute a distinct phenotype that is more than just a more severe form of SCZ. TRS is heritable, and genetics has been shown to play an important role in modulating response to antipsychotics. Important efforts have been put into place in order to better understand the genetic architecture of TRS, with the main goal of identifying reliable predictive markers that might improve the management and quality of life of TRS patients. However, the number of candidate gene and genome-wide association studies specifically focused on TRS is limited, and to date, findings do not allow the disentanglement of its polygenic nature. More recent studies implemented polygenic risk score, gene-based and machine learning methods to explore the genetics of TRS, reporting promising findings. In this review, we present an overview on the genetics of TRS, particularly focusing our discussion on studies implementing polygenic approaches.

Predictive modeling of schizophrenia from genomic data: Comparison of polygenic risk score with kernel support vector machines approach

A major controversy in psychiatric genetics is whether nonadditive genetic interaction effects contribute to the risk of highly polygenic disorders. We applied a support vector machines (SVMs) approach, which is capable of building linear and nonlinear models using kernel methods, to classify cases from controls in a large schizophrenia case-control sample of 11,853 subjects (5,554 cases and 6,299 controls) and compared its prediction accuracy with the polygenic risk score (PRS) approach. We also investigated whether SVMs are a suitable approach to detecting nonlinear genetic effects, that is, interactions. We found that PRS provided more accurate case/control classification than either linear or nonlinear SVMs, and give a tentative explanation why PRS outperforms both multivariate regression and linear kernel SVMs. In addition, we observe that nonlinear kernel SVMs showed higher classification accuracy than linear SVMs when a large number of SNPs are entered into the model. We conclude that SVMs are a potential tool for assessing the presence of interactions, prior to searching for them explicitly.