Orthrus: Towards Evolutionary and Functional RNA Foundation Models

Orthrus is a mature RNA model for RNA property prediction. It uses a Mamba encoder backbone, a variant of state-space models specifically designed for long-sequence data, such as RNA. Two versions of Orthrus are available: 4-track base version: Encodes the mRNA sequence with a simplified one-hot approach. 6-track large version: Adds biological context by including splice site indicators and coding sequence markers, which is crucial for accurate mRNA property prediction such as RNA half-life, ribosome load, and exon junction detection. This repository contains the annotations used to train Orthrus, as well as processed datasets used to evaluate Orthrus's ability to perform RNA property prediction. The datasets are taken from the following sources: Protein Subcellular Localization: Thul, P. J. et al. A subcellular map of the human proteome. Science 356 (2017). Mean Ribosome Load: Sugimoto, Y. & Ratcliffe, P. J. Isoform-resolved mRNA profiling of ribosome load defines interplay of HIF and mTOR dysregulation in kidney cancer. Nature Structural Molecular Biology 29, 871–880 (2022). RNA Halflife: Agarwal, V. & Kelley, D. R. The genetic and biochemical determinants of mRNA degradation rates in mammals. Genome Biol 23, 245 (2022). GO Molecular Function: Consortium, T. G. O. et al. The Gene Ontology knowledgebase in 2023. Genetics 224, iyad031 (2023). Dataset files encode sequence using a one-hot encoding using the vocabulary: [A, C, G, T].