AI Tool Identifies Gene Combinations Causing Complex Illnesses

Traditional methods such as genome-wide association studies often fall short in identifying how multiple genes interact to influence these traits, primarily due to limited statistical power and the vast number of possible gene combinations. Now, researchers have developed a new computational tool that uses generative artificial intelligence (AI) to uncover the genetic networks driving complex human conditions.

The tool, known as Transcriptome-Wide Conditional Variational Auto-Encoder (TWAVE), was created by a team of biophysicists at Northwestern University (Evanston, IL, USA) to improve understanding of how combinations of genes contribute to disease development. This model represents a breakthrough in linking genotype to phenotype by shifting the analytical focus from individual genes to coordinated gene groups.

TWAVE uses a generative AI model that amplifies limited gene expression data to detect patterns of gene activity associated with disease states. It works by simulating both healthy and diseased cellular environments and identifying how specific changes in gene expression can alter those states. Through a combined machine learning and optimization framework, TWAVE pinpoints the gene changes most likely to transition a cell from healthy to diseased or vice versa.

By analyzing gene expression rather than DNA sequences, TWAVE circumvents patient privacy concerns and captures the influence of environmental factors that can modulate gene activity. This focus allows the model to provide a dynamic view of cellular function and better reflect the complexity of real-world disease mechanisms. In the study published in the PNAS, TWAVE was tested across a variety of complex diseases. It successfully identified relevant gene combinations, including some that had been missed by existing techniques.

Notably, the model revealed that the same disease could arise from different gene sets in different individuals, highlighting the potential for more personalized therapeutic strategies. By revealing gene networks instead of isolated markers, TWAVE offers a powerful tool for developing targeted treatments that consider the specific genetic drivers in each patient. It paves the way for new interventions designed to modify the collective gene expression patterns that underlie chronic and multifactorial diseases.

“Many diseases are determined by a combination of genes — not just one,” said Northwestern’s Adilson Motter, the study’s senior author. “You can compare a disease like cancer to an airplane crash. In most cases, multiple failures need to occur for a plane to crash, and different combinations of failures can lead to similar outcomes. This complicates the task of pinpointing the causes. Our model helps simplify things by identifying the key players and their collective influence.”