Eukaryotic mRNA 3′ UTRs encode regulatory signals that shape post-transcriptional regulation, RNA fate and gene expression. However, a foundation model that spans broad eukaryotic 3′ UTR sequence space while unifying prediction, interpretation and design within a single framework is still lacking. Here we present EukaUTR, a 3′ UTR-specific foundation model trained on a large-scale eukaryotic 3′ UTR sequence corpus spanning diverse evolutionary lineages. Across 13 prediction tasks of post-transcriptional regulation, RNA fate and expression output, EukaUTR showed strong transferability, achieving relative improvements of up to 27.45% over the strongest baselines. Extending these capabilities to sequence design, EukaUTR generated novel 3′ UTRs that captured natural sequence patterns and regulatory features and EukaUTR-Guide extracted stability signals from small sets of 3′ UTRs to guide editing toward enhanced predicted stability. Together, EukaUTR establishes a sequence-to-function-to-design framework that connects broad eukaryotic representation learning with transferable functional prediction and mRNA design.
Background/Objectives: The 3′ untranslated region (3′ UTR) integrates cleavage and polyadenylation signals, microRNA targets, RNA-binding-protein sites, and RNA secondary structure, but the organizational levels that remain conserved during long-term sequence evolution are poorly understood. Methods: We analyzed the ra...
A. Kulikov, E. A. Sivoplyas, Oleg E. Lazebny· Genes· 0 citations
Genetic variants in untranslated regions (UTRs) can alter post-transcriptional regulation, yet functional interpretation remains difficult because regulatory mechanisms differ between 5′UTR and 3′UTR contexts and are not fully captured by generic noncoding-variant scores. We developed UTR-FNex, an interpretable machi...
Ya-Ning Li, Fang Ge, Lei Pan et al.· Journal of Chemical Informat...· 0 citations
RNA plays central roles in regulating information flow and provides a versatile substrate for engineering biological functions. While large language models (LLMs) have transformed natural language processing and protein design, a general framework connecting RNA foundation models to functional sequence design remains l...
Shi-Yuan Chen, Neil R. Fernandes, Wei Vivian Li et al.· bioRxiv· 0 citations
EPIC, the Eukaryotic Promoter and transcription Initiation prediction Challenge, relies on understudied animals spanning three phyla: octopus, oyster, milkweed bug, Indian meal moth, and shark and is open to everyone.
Ilya E. Vorontsov, Nikita Gryzunov, Bayley R. McDonald et al.· bioRxiv· 0 citations
The Vertebrate Regulatory MOdule Detector (VRMOD) provides a universal coordinate reference system for regulatory sequences across 309 vertebrate genomes and enables genome-wide annotation of non-coding regulatory elements in any vertebrate species using genomic sequence alone.
Tássia Mangetti Gonçalves, Casey L. Stewart, Samantha D. Baxley et al.· NAR Genomics and Bioinformat...· 0 citations
This Mini Review summarizes seq2func modelling in plant regulatory genomics, discusses plant-specific challenges and the complementary value of observational and perturbation assays, and proposes evaluation frameworks based on biologically meaningful distribution shifts.
Ming-Fang Zhan, Rui-Gang Wu· Frontiers in Plant Science· 0 citations
Zero-knowledge proofs can verify computing results without rerunning workloads or exposing sensitive data. Magnetar helps bring them into the mainstream with support for standard programming languages, major speed gains, and quantum-resistant security. The post Proof-based general-purpose computing with Magnetar appeared first on Microsoft Research.
