Weighted Diversified Sampling for Efficient Data-Driven Single-Cell Gene-Gene Interaction Discovery

TL;DR

This paper introduces a weighted diversified sampling method combined with Transformer models to efficiently identify significant gene-gene interactions in single-cell data, reducing data requirements by 99% while maintaining high performance.

Contribution

The study presents a novel weighted diversified sampling algorithm that enhances data efficiency in gene interaction discovery using Transformer models.

Findings

Sampling 1% of data yields comparable results to full dataset

The sampling algorithm computes diversity scores in two passes

Efficient subset generation accelerates gene interaction analysis

Abstract

Gene-gene interactions play a crucial role in the manifestation of complex human diseases. Uncovering significant gene-gene interactions is a challenging task. Here, we present an innovative approach utilizing data-driven computational tools, leveraging an advanced Transformer model, to unearth noteworthy gene-gene interactions. Despite the efficacy of Transformer models, their parameter intensity presents a bottleneck in data ingestion, hindering data efficiency. To mitigate this, we introduce a novel weighted diversified sampling algorithm. This algorithm computes the diversity score of each data sample in just two passes of the dataset, facilitating efficient subset generation for interaction discovery. Our extensive experimentation demonstrates that by sampling a mere 1\% of the single-cell dataset, we achieve performance comparable to that of utilizing the entire dataset.