AI will not give us precision medicine.

Area of Science:

• Biomedical Informatics
• Computational Biology
• Genomics

Background:

• Human DNA sequencing promised medical revolution but faced limitations due to genotype-phenotype complexity.
• Precision medicine utilizes molecular and clinical 'omics' data for tailored treatments.
• Large, heterogeneous datasets necessitate advanced computational analyses.

Purpose of the Study:

• To critically evaluate the biological relevance of artificial neural networks in precision medicine.
• To advocate for an integrated approach combining computational analysis with clinical expertise.

Main Methods:

• Analysis of artificial neural network (ANN) principles and their application in bioinformatics.
• Review of current practices in precision medicine data analysis.
• Conceptual framework for integrating data-driven and expertise-driven approaches.

Main Results:

• Artificial neural networks lack direct biological basis, despite claims of simulating the human brain.
• Current computational methods alone are insufficient for analyzing complex patient data.
• The effectiveness of precision medicine relies on bridging computational analysis and clinical insight.

Conclusions:

• Artificial neural networks are mathematical tools, not biological models, in the context of precision medicine.
• Effective precision medicine requires a synergistic collaboration between data scientists and clinicians.
• Understanding individual patient uniqueness is paramount and necessitates integrating diverse data types with expert knowledge.