Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

DNABERT-2: Efficient Foundation Model and Benchmark For Multi-Species Genome

About

Decoding the linguistic intricacies of the genome is a crucial problem in biology, and pre-trained foundational models such as DNABERT and Nucleotide Transformer have made significant strides in this area. Existing works have largely hinged on k-mer, fixed-length permutations of A, T, C, and G, as the token of the genome language due to its simplicity. However, we argue that the computation and sample inefficiencies introduced by k-mer tokenization are primary obstacles in developing large genome foundational models. We provide conceptual and empirical insights into genome tokenization, building on which we propose to replace k-mer tokenization with Byte Pair Encoding (BPE), a statistics-based data compression algorithm that constructs tokens by iteratively merging the most frequent co-occurring genome segment in the corpus. We demonstrate that BPE not only overcomes the limitations of k-mer tokenization but also benefits from the computational efficiency of non-overlapping tokenization. Based on these insights, we introduce DNABERT-2, a refined genome foundation model that adapts an efficient tokenizer and employs multiple strategies to overcome input length constraints, reduce time and memory expenditure, and enhance model capability. Furthermore, we identify the absence of a comprehensive and standardized benchmark for genome understanding as another significant impediment to fair comparative analysis. In response, we propose the Genome Understanding Evaluation (GUE), a comprehensive multi-species genome classification dataset that amalgamates $36$ distinct datasets across $9$ tasks, with input lengths ranging from $70$ to $10000$. Through comprehensive experiments on the GUE benchmark, we demonstrate that DNABERT-2 achieves comparable performance to the state-of-the-art model with $21 \times$ fewer parameters and approximately $92 \times$ less GPU time in pre-training.

Zhihan Zhou, Yanrong Ji, Weijian Li, Pratik Dutta, Ramana Davuluri, Han Liu• 2023

Related benchmarks

TaskDatasetResultRank
Genomic UnderstandingGUE
H3 Score81.6
14
Genomic UnderstandingGUE (test)
Average Score67.77
10
Core Promoter DetectionCPD
Score (all)0.6937
8
Splice Site PredictionSSP
Reconstruction0.8499
8
Transcription Factor PredictionTFP
Performance Score 071.99
8
Promoter DetectionPD
Score (all)86.77
8
Core Promoter DetectionGUE--
5
Epigenetic Marks PredictionGUE
H378.27
4
Splice Site PredictionGUE
Reconstruct0.8499
4
Species ClassificationGUE+ Species Classification (SC) v1 (test)
Accuracy (Fungi)93.04
3
Showing 10 of 17 rows

Other info

Code

Follow for update