Some statistics in bioinformatics: the fifth Armitage Lecture

Area of Science:

• Biostatistics
• Bioinformatics
• Epidemiology

Background:

• Peter Armitage's pioneering research in sequential methods for medical trials and survival curve comparison.
• The enduring influence of Armitage's early work on subsequent statistical methodologies.
• The evolving relationship between biological sciences and statistical analysis.

Discussion:

• The paper presents several statistical 'truths' illustrated by biological morphology, suggesting biology's ascendance.
• A cautionary tale of a proteomics ovarian cancer study highlights pitfalls when bioinformatics and epidemiology intersect poorly, especially with flawed study design.
• A successful application of statistical bioinformatics in gene profiling for mouse embryonic stem cell development is presented.

Key Insights:

• Biology's growing dominance over traditional statistical approaches in scientific research.
• The critical importance of robust study design in bioinformatics and epidemiological research.
• Gene profiling as a powerful tool in understanding developmental biology.

Outlook:

• Future research directions integrating advanced bioinformatics with epidemiological studies.
• The need for improved methodologies at the intersection of biology, statistics, and data science.
• Continued exploration of gene networks in developmental processes.