AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

TrinityDNA: A Bio-Inspired Foundational Model for Efficient Long-Sequence DNA Modeling

Qirong Yang, Yucheng Guo, Zicheng Liu, Yujie Yang, Qijin Yin, Siyuan Li, Shaomin Ji, Linlin Chao, Xiaoming Zhang

AAAI 2026 · Architectures and efficiency › Attention and transformers · approach: State space and recurrent sequence models · 2 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add Gated Reverse Complement (ablation)TrinityDNA base · Pre-training perplexityperplexity: 2.731 -> 2.599Improved
add Groove Fusion Module (ablation)TrinityDNA with GRC · Pre-training perplexityperplexity: 2.599 -> 2.534Improved
add Sliding Multi-Window Attention (ablation)TrinityDNA with GRC+GFM · Pre-training perplexityperplexity: 2.534 -> 2.544Hurt
add Sliding Multi-Window Attention (ablation)TrinityDNA with GRC+GFM · Pre-training computeTFLOPs: 64.5 -> 44.5 (-31%)Improved
evolutionary training: prokaryote pretraining then multi-species vs training from scratch on combined data (ablation)TrinityDNA · Multi-species perplexityperplexityImproved
increase context length from 8k to 100k (ablation)TrinityDNA · Eukaryotic perplexityperplexityImproved
post-train on eukaryotic data (TrinityDNA vs TrinityMicroDNA)TrinityMicroDNA-1B · Zero-shot eukaryote averageSpearman: 0.404 -> 0.699Improved
post-train on eukaryotic data (TrinityDNA vs TrinityMicroDNA)TrinityMicroDNA-1B · Zero-shot prokaryote averageSpearman: 0.475 -> 0.366Hurt
use TrinityDNA 1B vs larger EVO2 40BTrinityDNA · Zero-shot eukaryote averageSpearman: 0.667 -> 0.699Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy