From small studies to precision medicine: prioritizing candidate biomarkers

Area of Science:

• Biostatistics
• Genomics
• Bioinformatics

Background:

• Personalized medicine and big data present challenges in biomarker development.
• Best practices for extracting candidate biomarkers from small, hypothesis-generating studies remain unclear.

Purpose of the Study:

• To compare methods for identifying patient-specific aberrant gene expression events in small- to medium-sized studies.
• To determine the optimal approach for biomarker discovery in the context of precision medicine.

Main Methods:

• Evaluation of various data-analytic methods for detecting aberrant expression.
• Focus on outlying degree methods for identifying significant expression events.
• Analysis of studies with sample sizes ranging from 10 to 50.

Main Results:

• Outlying degree methods demonstrated superior performance in detecting patient-specific aberrant expression events.
• These methods proved effective even with limited sample sizes (10-50 samples).
• The findings support the utility of outlying degree methods for hypothesis-generating biomarker studies.

Conclusions:

• Outlying degree methods represent a best practice for biomarker discovery in small-scale studies.
• These methods are crucial for advancing personalized medicine by identifying relevant biomarkers.
• Further research should leverage these techniques for robust biomarker development.