Researchers at the Centro Nacional de Análisis Genómico (CNAG), located in the Barcelona Science Park, have developed Pheno-Ranker, an open-source toolkit that transforms complex phenotypic data into actionable information, accelerating the discovery of diagnoses for rare diseases and facilitating the creation of cohorts for cross-disease analyses. The study, published in BMC Bioinformatics, is fully accessible and helps researchers classify patients and find similar cases with just a few clicks, processing various phenotypic data formats and applying semantic similarity algorithms.

Understanding phenotypic data, that is, the observable characteristics we see in individuals, is fundamental for advancing clinical research. These data provide crucial insights into how diseases manifest and progress, enabling scientists and clinicians to make more accurate diagnoses, classify patients based on their unique traits, and develop personalised treatment plans. However, analysing and comparing phenotypic information across different patients and cohorts is a complex and time-consuming task.

Addressing this challenge, the Biomedical Genomics Group at CNAG, led by its director Dr Ivo Gut, has made a significant breakthrough with the launch of Pheno-Ranker. This novel, open-source toolkit is designed specifically for the scientific community to facilitate the comparison and analysis of phenotypic data. The project was developed by CNAG Biomedical Informatics researchers Manuel Rueda and Ivo Leist, in collaboration with Pfizer and the University of Granada’s Junta de Andalucía Centre for Genomics and Oncological Research, Granada (Spain) in the course of the 3TR project.

Pheno-Ranker simplifies and streamlines the process of making sense of complex phenotypic profiles. The toolkit is flexible and easy to use: it accepts input files in widely used formats such as JSON, TSV and CSV, which can include free text or structured terms from ontologies such as the Human Phenotype Ontology (HPO). It processes these data through semantic similarity algorithms, allowing researchers to quantify the similarity between individuals and between patients and reference cohorts. The output is presented in clear, ranked lists of similarities, which greatly simplifies the identification of relevant cases, potential diagnoses, and clinically meaningful patterns.

The software offers two complementary modes of analysis that help researchers explore phenotypic data more effectively. In cohort mode, the toolkit allows scientists to analyse groups of individuals as a whole. This helps to identify patterns, classify patient subgroups, or detect potential outliers within a cohort. In patient mode, the focus shifts to the individual level. Researchers can compare a single patient’s phenotypic profile against an entire cohort, a function especially useful in rare disease research. This mode supports the identification of similar cases and can assist in generating potential diagnoses based on phenotypic similarities. The toolkit’s powerful web-based interface streamlines data calculations and supports direct visual analytics.

“A common issue is the need to cluster patients by similarity or identify the closest match for a given patient. That’s why we developed Pheno-Ranker — to allow users, for instance, to find the closest matches for rare disease (RD) patients in resources like OMIM or Orphanet, which together catalog thousands of rare diseases. Unlike a general GenAI or web search, Pheno-Ranker provides statistically meaningful results, making it a more reliable tool for clinical interpretation and decision-making”, explains one of the authors of the study and Bioinformatician at the Biomedical Genomics Group at CNAG.

By being open-source, Pheno-Ranker promotes wide adoption and collaboration within the research community. This fosters innovation and accelerates discoveries in clinical genomics and rare disease research worldwide. With Pheno-Ranker, CNAG is contributing a powerful new tool to help translate phenotypic data into actionable insights, ultimately improving patient diagnosis and care.

» Article of reference: Leist, Ivo C., et al. «Pheno-Ranker: a toolkit for comparison of phenotypic data stored in GA4GH standards and beyond». BMC Bioinformatics, vol. 25, n.o 1, diciembre de 2024, p. 373. BioMed Central, doi: https://doi.org/10.1186/s12859-024-05993-2.

» Link to the news: CNAG website [+]