Surveying Multi-Agent Genomic Systems With A Spotlight On Oncology

Genomic analysis of a patient’s data has become essential for understanding their diseases and proposing the necessary treatments. Traditional genomic analysis tools are incapable of reliably transforming raw sequencing results into information that clinicians can effectively use. To address these challenges, researchers have invested significant effort in developing multi-agent systems. These systems enable the use of indepen- dent agents that collaborate to perform well-defined tasks. These agents can leverage advances in artificial intelligence, such as large language models, which add a cognitive layer and enhance decision support. The objective of this research is to study multi-agent systems developed in the literature and to compare them based on well-defined criteria in order to identify current shortcomings and understand the gaps within this research field. In this review, we evaluate 25 genomic multi-agent systems published between 2022 and 2025. We classify these systems according to their application domain, identify their coverage of the ten main stages of the analysis pipeline, and characterize four clinical functions: personalized medicine, therapeutic recommendations, interoperability standards, and explainability. The analysis reveals a strong orientation toward general genomics, limited consideration of the early stages of the pipeline, and very restricted advanced capabilities regarding clinical functionalities—an observation worth noting. This situation indicates the need to develop multi-agent systems that ensure automation across a broader range of processes and incorporate more clinically oriented functionalities.