A support vector machine model provides an accurate transcript-level-based diagnostic for major depressive disorder.

E E Redei,A Y Xue,N Bagheri,J S Yu

doi:10.1038/tp.2016.198

E E Redei, A Y Xue + Show 2 more

Open Access

https://doi.org/10.1038/tp.2016.198

Copy DOI

Abstract

Major depressive disorder (MDD) is a critical cause of morbidity and disability with an economic cost of hundreds of billions of dollars each year, necessitating more effective treatment strategies and novel approaches to translational research. A notable barrier in addressing this public health threat involves reliable identification of the disorder, as many affected individuals remain undiagnosed or misdiagnosed. An objective blood-based diagnostic test using transcript levels of a panel of markers would provide an invaluable tool for MDD as the infrastructure—including equipment, trained personnel, billing, and governmental approval—for similar tests is well established in clinics worldwide. Here we present a supervised classification model utilizing support vector machines (SVMs) for the analysis of transcriptomic data readily obtained from a peripheral blood specimen. The model was trained on data from subjects with MDD (n=32) and age- and gender-matched controls (n=32). This SVM model provides a cross-validated sensitivity and specificity of 90.6% for the diagnosis of MDD using a panel of 10 transcripts. We applied a logistic equation on the SVM model and quantified a likelihood of depression score. This score gives the probability of a MDD diagnosis and allows the tuning of specificity and sensitivity for individual patients to bring personalized medicine closer in psychiatry.

Highlights

Undiagnosed cases of major depressive disorder (MDD) pose a major detriment to society by contributing to disability,[1] comorbid health conditions,[2] and, in many cases, suicide.[3]
It has been suggested previously that inborn or acquired errors of metabolism are often accompanied by psychiatric symptoms, whether as a consequence or a cause, only future research can tell.[36]. These molecular characterizations suggest that there may be multiple converging causes that lead to depression through altering the availability of proteins, perhaps enzymes, central to the etiology of Major depressive disorder (MDD). This is the first report to our knowledge that describes in complete detail different classification models for future diagnostic purposes built using a panel of transcript abundances
The logistic regression and support vector machines (SVMs) models offer high sensitivity and specificity in predicting subjects with MDD. Both models have their advantages: logistic regression can assign a probability of a subject having MDD and the SVM model has fewer explanatory variables and is less sensitive to outliers

Summary

Introduction

Undiagnosed cases of major depressive disorder (MDD) pose a major detriment to society by contributing to disability,[1] comorbid health conditions,[2] and, in many cases, suicide.[3] Approximately. 16.2% of the US population suffers from MDD at least once in a lifetime,[2] with significant percentages left undiagnosed, misdiagnosed, and/or untreated.[4] Individuals with MDD are identified by self-reported changes in behavior, mood, and clinical examination.[5] certain subpopulations—such as children,[6] adolescents,[7] and elderly individuals8—prove difficult to diagnose with these methods. This difficulty may be attributed to comorbid mood disorders, trouble in communicating with doctors,[9] and unwillingness to seek clinical help because of social stigma.[10]

Methods

Results

Conclusion