Ribonucleic acid (RNA) plays fundamental roles in biological systems, from carrying genetic information to performing enzymatic function. Understanding and designing RNA can enable novel therapeutic application and biotechnological innovation. To enhance RNA design, in this paper we introduce RiboGen, the first deep learning model to simultaneously generate RNA sequence and all-atom 3D structure. RiboGen leverages the standard Flow Matching with Discrete Flow Matching in a multimodal data representation. RiboGen is based on Euclidean Equivariant neural networks for efficiently processing and learning three-dimensional geometry. Our experiments show that RiboGen can efficiently generate chemically plausible and self-consistent RNA samples. Our results suggest that co-generation of sequence and structure is a competitive approach for modeling RNA.

Mapping the conformational dynamics of proteins is crucial for elucidating their functional mechanisms. While Molecular Dynamics (MD) simulation enables detailed time evolution of protein motion, its computational toll hinders its use in practice. To address this challenge, multiple deep learning models for reproducing and accelerating MD have been proposed drawing on transport-based generative methods. However, existing work focuses on generation through transport of samples from prior distributions, that can often be distant from the data manifold. The recently proposed framework of stochastic interpolants, instead, enables transport between arbitrary distribution endpoints. Building upon this work, we introduce EquiJump, a transferable SO(3)-equivariant model that bridges all-atom protein dynamics simulation time steps directly. Our approach unifies diverse sampling methods and is benchmarked against existing models on trajectory data of fast folding proteins. EquiJump achieves state-of-the-art results on dynamics simulation with a transferable model on all of the fast folding proteins. Proteins are the workhorses of the cell, and simulating their dynamics is critical to biological discovery and drug design (Karplus and Kuriyan, 2005).

Electron density prediction stands as a cornerstone challenge in molecular systems, pivotal for various applications such as understanding molecular interactions and conducting precise quantum mechanical calculations. However, the scaling of density functional theory (DFT) calculations is prohibitively expensive. Machine learning methods provide an alternative, offering efficiency and accuracy. We introduce a novel SE(3)-equivariant architecture, drawing inspiration from Slater-Type Orbitals (STO), to learn representations of molecular electronic structures. Our approach offers an alternative functional form for learned orbital-like molecular representation. We showcase the effectiveness of our method by achieving SOTA prediction accuracy of molecular electron density with 30-70\% improvement over other work on Molecular Dynamics data.

Three-dimensional native states of natural proteins display recurring and hierarchical patterns. Yet, traditional graph-based modeling of protein structures is often limited to operate within a single fine-grained resolution, and lacks hourglass neural architectures to learn those high-level building blocks. We narrow this gap by introducing Ophiuchus, an SO(3)-equivariant coarse-graining model that efficiently operates on all-atom protein structures. Our model departs from current approaches that employ graph modeling, instead focusing on local convolutional coarsening to model sequence-motif interactions with efficient time complexity in protein length. We measure the reconstruction capabilities of Ophiuchus across different compression rates, and compare it to existing models. We examine the learned latent space and demonstrate its utility through conformational interpolation. Our experiments demonstrate Ophiuchus to be a scalable basis for efficient protein modeling and generation. Proteins form the basis of all biological processes and understanding them is critical to biological discovery, medical research and drug development. Their three-dimensional structures often display modular organization across multiple scales, making them promising candidates for modeling in motif-based design spaces [Bystroff & Baker (1998); Mackenzie & Grigoryan (2017); Swanson et al. (2022)]. Harnessing these coarser, lower-frequency building blocks is of great relevance to the investigation of the mechanisms behind protein evolution, folding and dynamics [Mackenzie et al. (2016)], and may be instrumental in enabling more efficient computation on protein structural data through coarse and latent variable modeling [Kmiecik et al. (2016); Ramaswamy et al. (2021)]. Recent developments in deep learning architectures applied to protein sequences and structures demonstrate the remarkable capabilities of neural models in the domain of protein modeling and design [Jumper et al. (2021); Baek et al. (2021b); Ingraham et al. (2022); Watson et al. (2022)].

A key factor in the modern success of deep learning is the astonishing expressive power of neural networks. However, this comes at the cost of complex, black-boxed models that are unable to extrapolate beyond the domain of the training dataset, conflicting with goals of expressing physical laws or building human-readable programs. In this paper, we introduce OccamNet, a neural network model that can find interpretable, compact and sparse solutions for fitting data, \`{a} la Occam's razor. Our model defines a probability distribution over a non-differentiable function space, and we introduce an optimization method that samples functions and updates the weights based on cross-entropy matching in an evolutionary strategy: we train by biasing the probability mass towards better fitting solutions. We demonstrate that we can fit a variety of algorithms, ranging from simple analytic functions through recursive programs to even simple image classification. Our method takes minimal memory footprint, does not require AI accelerators for efficient training, fits complicated functions in minutes of training on a single CPU, and demonstrates significant performance gains when scaled on GPU. Our implementation, demonstrations and instructions for reproducing the experiments are available at https://github.com/AllanSCosta/occam-net.