Combining gene expression, demographic and clinical data in modeling disease: a case study of bipolar disorder and schizophrenia
2008

Modeling Disease with Gene Expression and Clinical Data

Sample size: 332 publication Evidence: high

Author Information

Author(s): Jan Struyf, Seth Dobrin, David Page

Primary Institution: Katholieke Universiteit Leuven

Hypothesis

Can classification algorithms improve by including demographic and clinical data alongside gene expression data in distinguishing bipolar disorder and schizophrenia from control?

Conclusion

Support vector machines can effectively distinguish bipolar disorder and schizophrenia from normal controls, with improved accuracy when demographic and clinical data are included.

Supporting Evidence

  • SVMs achieved an AUC of 0.92 for bipolar disorder versus control and 0.91 for schizophrenia versus control using gene expression data alone.
  • Including demographic and clinical data improved AUC to 0.97 for bipolar disorder and 0.94 for schizophrenia.
  • Support vector machines outperformed other classification algorithms in distinguishing between the diseases and controls.

Takeaway

Scientists found that using both gene data and information about people's backgrounds helps better identify mental health issues like bipolar disorder and schizophrenia.

Methodology

The study compared six classification algorithms using gene expression data and demographic/clinical data to distinguish between bipolar disorder, schizophrenia, and control groups.

Potential Biases

Potential bias due to the influence of alcohol and drug use on gene expression data.

Limitations

The study is retrospective and may be affected by confounding variables such as alcohol and drug use.

Participant Demographics

The study included 115 schizophrenia patients, 105 bipolar disorder patients, and 112 controls, with demographic data on age, sex, and substance use.

Statistical Information

P-Value

p<0.05

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1186/1471-2164-9-531

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication