Conditional Monte Carlo Tree Diffusion for Designing Cell-Type-Specific and Biologically Faithful Regulatory DNA

TL;DR

DNA-CRAFT is a novel generative framework combining class-conditioned diffusion and Monte Carlo tree search to design cell-type-specific regulatory DNA with high biological fidelity.

Contribution

It introduces a new method that integrates class-conditioned diffusion models with Monte Carlo tree search for precise regulatory DNA design.

Findings

Achieves high predicted cell-type-specific activity.

Demonstrates superior trade-offs over existing methods.

Validates effectiveness on human and immune cell regulatory sequences.

Abstract

Designing regulatory DNA elements with precise cell-type-specific activity is broadly relevant for cell engineering and gene therapy. Deep generative models can generate functional gene-regulatory elements, but existing methods struggle to achieve high specificity against undesired cell types while adhering to the genome's natural regulatory grammar. Here, we introduce DNA-CRAFT, a generative framework that integrates class-conditioned discrete diffusion with Monte Carlo tree search to design cell-type-specific and biologically faithful regulatory elements. We first train a discrete diffusion model on the ENCODE registry of 3.2 million candidate regulatory elements. Second, we condition the model to learn class-specific regulatory grammars of naturally occurring DNA sequences, including enhancers and promoters. Third, we employ conditional Monte Carlo tree guidance, an inference-time alignment algorithm designed to maximize the differential regulatory activity between desired and undesired cell types. By benchmarking DNA-CRAFT on regulatory sequence design tasks for human cell lines and immune cell types, we demonstrate that our model generates sequences with high predicted cell-type-specific activity and biological fidelity, achieving the best trade-offs compared to methods that use diffusion, autoregressive models, and gradient-based optimization.