# Revisiting Minamata disease through computational phenotypic similarity analysis

## TL;DR

This study uses modern computational methods to analyze the symptoms of Minamata disease and find connections to other neurological disorders.

## Contribution

The study introduces a network-based approach to analyze Minamata disease's phenotypic profile and compare it with thousands of other diseases.

## Key findings

- Minamata disease shows strong phenotypic similarity to movement and neurodegenerative disorders like cyanide-induced parkinsonism.
- Computational analysis revealed a robust consensus network of diseases with overlapping symptoms.
- The study demonstrates how historical data can be integrated with modern tools to explore environmental disease mechanisms.

## Abstract

Minamata disease, a severe neurological disorder caused by methylmercury exposure in 1950s Japan, is historically recognized for its profound impact on environmental health awareness. However, its phenotypic complexity and potential overlap with other neurological disorders have not been systematically assessed in a modern computational framework. In this study, we adopt a network approach to reinterpret Minamata disease within a broader disease similarity landscape. We mapped clinical symptoms from an extensive epidemiological survey of 269 Minamata patients to standardized Human Phenotype Ontology (HPO) terms, constructing a comprehensive phenotypic profile. Using network-based and computational similarity measures—Jaccard Index, ontology-informed metrics (Resnik and GraphIC), and information retrieval techniques (TF-IDF with query expansion), we compared this profile to over 12,000 diseases. Our results consistently identified strong phenotypic ties between Minamata disease and several movement and neurodegenerative disorders, including cyanide-induced parkinsonism and progressive supranuclear palsy. A weighted rank aggregation across methods revealed a robust consensus network of diseases with overlapping symptomatology, underscoring the systemic nature of these complex neurological disorders. Our study highlights the utility of integrating historical epidemiological data with contemporary network tools to reveal novel associations between environmental exposures and systemic pathophysiological responses. Our findings provide a blueprint for exploring environmentally triggered disease mechanisms and their broader implications for network-based understanding of human disease.