Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Hou, Yusen, Long, Weicai, Hu, Haitao, Su, Houcheng, Feng, Junning, Zhang, Yanlin
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2604.05775
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866913010409275392
author Hou, Yusen
Long, Weicai
Hu, Haitao
Su, Houcheng
Feng, Junning
Zhang, Yanlin
author_facet Hou, Yusen
Long, Weicai
Hu, Haitao
Su, Houcheng
Feng, Junning
Zhang, Yanlin
contents Bacteriophages, often referred to as the dark matter of the biosphere, play a critical role in regulating microbial ecosystems and in antibiotic alternatives. Thus, accurate interpretation of their genomes holds significant scientific and practical value. While general-purpose Large Language Models (LLMs) excel at understanding biological texts, their ability to directly interpret raw nucleotide sequences and perform biological reasoning remains underexplored. To address this, we introduce PhageBench, the first benchmark designed to evaluate phage genome understanding by mirroring the workflow of bioinformatics experts. The dataset contains 5,600 high-quality samples covering five core tasks across three stages: Screening, Quality Control, and Phenotype Annotation. Our evaluation of eight LLMs reveals that general-purpose reasoning models significantly outperform random baselines in phage contig identification and host prediction, demonstrating promising potential for genomic understanding. However, they exhibit significant limitations in complex reasoning tasks involving long-range dependencies and fine-grained functional localization. These findings highlight the necessity of developing next-generation models with enhanced reasoning capabilities for biological sequences.
format Preprint
id arxiv_https___arxiv_org_abs_2604_05775
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle PhageBench: Can LLMs Understand Raw Bacteriophage Genomes?
Hou, Yusen
Long, Weicai
Hu, Haitao
Su, Houcheng
Feng, Junning
Zhang, Yanlin
Computation and Language
Genomics
Bacteriophages, often referred to as the dark matter of the biosphere, play a critical role in regulating microbial ecosystems and in antibiotic alternatives. Thus, accurate interpretation of their genomes holds significant scientific and practical value. While general-purpose Large Language Models (LLMs) excel at understanding biological texts, their ability to directly interpret raw nucleotide sequences and perform biological reasoning remains underexplored. To address this, we introduce PhageBench, the first benchmark designed to evaluate phage genome understanding by mirroring the workflow of bioinformatics experts. The dataset contains 5,600 high-quality samples covering five core tasks across three stages: Screening, Quality Control, and Phenotype Annotation. Our evaluation of eight LLMs reveals that general-purpose reasoning models significantly outperform random baselines in phage contig identification and host prediction, demonstrating promising potential for genomic understanding. However, they exhibit significant limitations in complex reasoning tasks involving long-range dependencies and fine-grained functional localization. These findings highlight the necessity of developing next-generation models with enhanced reasoning capabilities for biological sequences.
title PhageBench: Can LLMs Understand Raw Bacteriophage Genomes?
topic Computation and Language
Genomics
url https://arxiv.org/abs/2604.05775