SimNetX: tinkering with patient similarity networks to understand biomedical data

Analyzing complex biomedical data often requires statistical and machine learning expertise, creating barriers for clinicians, laboratory scientists, and other non-technical users. Patient similarity networks (PSNs) offer an intuitive way to explore patient relationships and patterns, making data interpretation more accessible. However, constructing and analyzing PSNs typically involves multiple software tools or programming skills, limiting their usability for those without technical expertise. In this article, we introduce an approach that enables non-technical users to analyze biomedical data through PSNs without requiring programming knowledge. By integrating key functionalities-such as transforming vector-based data into networks, interactively exploring patient relationships, and applying statistical insights-this approach bridges the gap between complex data analysis and domain experts. To facilitate this, we provide a tool designed to implement these methods in an intuitive, interactive environment. We demonstrate its practical application using two well-known datasets as well as two real-world biomedical datasets, showing how non-experts can generate hypotheses and extract meaningful insights through visual exploration and built-in simple statistical analysis.