Premium
Blood‐based dynamic genomic signature for obsessive–compulsive disorder
Author(s) -
Wang Yuan,
Cheng Changming,
Zhang Zongfeng,
Wang Jianyu,
Wang Yao,
Li Xiaoping,
Gao Rui,
Wang Zhen,
Fang Yiru,
Wang Jijun,
Wang Min,
Fan Qing,
Periya Sanggetha,
Zhang Haiyin,
Tsuang Ming T.,
Liew ChoongChin
Publication year - 2018
Publication title -
american journal of medical genetics part b: neuropsychiatric genetics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.393
H-Index - 126
eISSN - 1552-485X
pISSN - 1552-4841
DOI - 10.1002/ajmg.b.32675
Subject(s) - schizophrenia (object oriented programming) , major depressive disorder , receiver operating characteristic , psychiatry , cohort , logistic regression , medicine , disease , clinical psychology , cognition
No biologically based diagnostic criteria are in clinical use today for obsessive–compulsive disorder (OCD), schizophrenia, and major depressive disorder (MDD), which are defined with reference to Diagnostic and Statistical Manual clinical symptoms alone. However, these disorders cannot always be well distinguished on clinical grounds and may also be comorbid. A biological blood‐based dynamic genomic signature that can differentiate among OCD, MDD, and schizophrenia would therefore be of great utility. This study enrolled 77 patients with OCD, 67 controls with no psychiatric illness, 39 patients with MDD, and 40 with schizophrenia. An OCD‐specific gene signature was identified using blood gene expression analysis to construct a predictive model of OCD that can differentiate this disorder from healthy controls, MDD, and schizophrenia using a logistic regression algorithm. To verify that the genes selected were not derived as a result of chance, the algorithm was tested twice. First, the algorithm was used to predict the cohort with true disease/control status and second, the algorithm predicted the cohort with disease/control status randomly reassigned (null set). A six‐gene panel ( COPS7A, FKBP1A, FIBP, TP73‐AS1, SDF4, and GOLGA8A ) discriminated patients with OCD from healthy controls, MDD, and schizophrenia in the training set (with an area under the receiver‐operating‐characteristic curve of 0.938; accuracy, 86%; sensitivity, 88%; and specificity, 85%). Our findings indicate that a blood transcriptomic signature can distinguish OCD from healthy controls, MDD, and schizophrenia. This finding further confirms the feasibility of using dynamic blood‐based genomic signatures in psychiatric disorders and may provide a useful tool for clinical staff engaged in OCD diagnosis and decision making.