Researchers Launched Global Gene Expression Challenge
The new EPIC competition tasks teams with predicting gene expression in non-model animal species using only DNA sequences.
Updated on Sept. 24, 2026 in Life Sciences
Researchers have announced the Eukaryotic Promoter and transcription Initiation prediction Challenge (EPIC), an open competition designed to improve models for non-model animal species. The project focuses on predicting gene expression in octopus, oyster, milkweed bug, Indian meal moth, and shark species.
Why it matters
Current gene expression models lack sufficient benchmarks for generalization beyond humans and primary model species. By requiring participants to work across diverse animal lineages, this challenge aims to refine predictive accuracy for genomics outside of traditional laboratory organisms.
Teams are provided with strand-specific, single-nucleotide-resolution initiation profiles covering 80-95% of the genome to train their models. To qualify for formal consortium authorship, entries must exceed the established dinucleotide precision baseline.
The players
Eukaryotic Promoter and transcription Initiation prediction Challenge
A global research initiative focused on benchmarking computational models for gene expression prediction across non-traditional animal genomes.
The details
The competition requires participants to build prediction models that infer gene expression profiles for five specific non-model species using only the raw DNA sequence as an input. Researchers define initiation profiles as the specific genomic regions where transcription—the process where cells copy DNA into RNA—begins. By using a diverse set of animals, the challenge tests how well machine learning models generalize biological patterns across disparate evolutionary lineages.
Timeline
September 24, 2026: The EPIC challenge announcement was published.
December 31, 2026: The deadline for all EPIC prediction challenge submissions.
The Tech Race
The challenge seeks to bridge the gap between highly annotated human data and the lack of comparable benchmarks for other animal species. This effort follows a pattern set by early genomics projects to standardize how researchers interpret and map gene expression across diverse life forms.
The competition is open to the public, allowing researchers and hobbyists with expertise in genomics to contribute to the model development. The resulting predictive tools will likely be released for scientific use following the challenge, potentially assisting ecologists and evolutionary biologists.
The takeaway
The challenge serves as a primary testbed for how machine learning can interpret gene expression using only sequence data. Interested researchers should monitor the EPIC challenge scientific preprint for updates leading up to the December 31, 2026, deadline.
What happens next
Participants must submit their models by December 31, 2026, with successful teams invited to join as authors on the final peer-reviewed publication.
Further reading
For broader insights on computational biology, visit our Life Sciences section.
More information
View the EPIC challenge scientific preprint for full methodological details.
