# PanGIA: A universal framework for identifying association between ncRNAs and diseases

## TL;DR

PanGIA is a new framework that predicts how different types of noncoding RNAs are linked to diseases, outperforming existing methods and offering strong biological insights.

## Contribution

PanGIA introduces a novel computational framework that simultaneously predicts associations between multiple ncRNA types and diseases, leveraging cross-type interactions.

## Key findings

- PanGIA outperforms single-type state-of-the-art methods in predicting ncRNA–disease associations.
- The model remains robust even when nodes or ncRNA types are removed, confirming the value of cross-type information.
- Case studies validate the model's predictions with literature evidence, showing strong biological interpretability.

## Abstract

With the growing recognition of the important roles noncoding RNAs (ncRNAs) play in various biological functions, especially their potential involvement in many human diseases, predicting ncRNA–disease associations has become a key challenge in biomedical research.

Although many computational methods have been proposed to predict ncRNA–disease associations, most of these methods focus on a single type of ncRNA. However, the competitive and cooperative interactions among different types of ncRNAs are closely related to their functional roles in disease associations. To address this limitation, we propose a novel computational framework, PanGIA (Pan-ncRNA Graph-Interaction Attention network), designed to simultaneously predict potential associations between multiple types of noncoding RNAs, including microRNAs (miRNAs), long noncoding RNAs (lncRNAs), circular RNAs (circRNAs), and PIWI-interacting RNAs (piRNAs), and diseases. Experimental results show that PanGIA outperforms type-specific SOTA methods in both individual and comprehensive predictions. It remains robust even when nodes or ncRNA types are removed, and ablation studies confirm the benefits of cross-type information. PanGIA also outperforms several single-type state-of-the-art methods across multiple metrics.

PanGIA demonstrates significant advantages in predicting disease associations for different types of ncRNAs, including miRNAs, lncRNAs, circRNAs, and piRNAs. Case studies further confirm the accuracy of the model’s predictions, as all high-confidence associations were supported by literature evidence. This demonstrates the model’s strong biological interpretability and promising potential for practical applications. The successful application of PanGIA provides a new paradigm for exploring disease-associated ncRNAs, highlighting their immense potential in the field of biomedical research.