SIGN: similarity identification in gene expression.

Area of Science:

• Computational biology
• Genomics
• Bioinformatics

Background:

• High-throughput molecular profiles aid in stratifying phenotypes and identifying biological states.
• Pathway activities are utilized as biological features in machine learning schemes.

Purpose of the Study:

• To develop an open-source R package named SIGN (Similarity Identification in Gene expressioN).
• To facilitate the use of pathway activities and gene expression patterns for identifying similarities between biological samples.
• To introduce a new measure, the transcriptional similarity coefficient, for capturing gene expression pattern similarity within biological pathways.

Main Methods:

• Development of the SIGN R package.
• Definition of the transcriptional similarity coefficient.
• Application of SIGN to discriminate breast tumor subtypes and predict patient survival.

Main Results:

• SIGN effectively discriminates breast tumor subtypes and predicts patient survival.
• SIGN outperforms existing models in predicting breast cancer patient survival, as demonstrated by DREAM challenge data.
• The transcriptional similarity coefficient captures gene expression pattern similarity within pathways.

Conclusions:

• SIGN serves as a valuable tool for interrogating pathway activity and gene expression patterns.
• SIGN enhances unsupervised and supervised learning schemes for improved prognostic risk estimation in cancer patients.
• The R package is available for the biomedical research community.