BaseCamp --- An Agentic AI Framework for Automating DNA Sequencing Data Pipelines

BaseCamp：用于自动化 DNA 测序数据流水线的智能体 AI 框架

Abstract: DNA sequencing pipelines, spanning quality control, alignment, variant calling, and annotation, are now reliably executed by workflow management systems that orchestrate established bioinformatics tools at scale.

摘要： DNA 测序流水线涵盖了质量控制、比对、变异调用和注释等环节，目前已能通过工作流管理系统可靠地大规模执行现有的生物信息学工具。

What remains manual is the decision layer surrounding that execution: selecting quality thresholds appropriate to a sample and platform, adjudicating borderline variant calls, diagnosing anomalies, and determining which findings warrant expert review. These decisions are repetitive, judgment-intensive, inconsistent across operators, and frequently undocumented.

目前仍需人工操作的是执行过程中的决策层：包括选择适合特定样本和平台的质量阈值、判定临界变异调用、诊断异常情况，以及确定哪些发现需要专家审查。这些决策具有重复性、高度依赖判断力、在不同操作员之间不一致，且往往缺乏记录。

This paper introduces BaseCamp, a novel agentic AI framework for automating the decision layer of DNA sequencing pipelines. The framework decomposes the pipeline into six specialized AI agents, covering sample intake and quality control, alignment, variant calling, annotation, cross-stage monitoring, and reporting.

本文介绍了 BaseCamp，这是一个用于自动化 DNA 测序流水线决策层的新型智能体 AI 框架。该框架将流水线分解为六个专门的 AI 智能体，涵盖样本接收与质量控制、比对、变异调用、注释、跨阶段监控以及报告生成。

Critically, BaseCamp agents do not perform sequence analysis: established tools execute alignment, calling, and annotation, while the agents select among them, configure them, interpret their output, and decide what follows. This confines language model reasoning to the judgment layer where it is reliable and preserves the reproducibility existing tooling guarantees.

关键在于，BaseCamp 智能体并不直接执行序列分析：现有的成熟工具负责执行比对、调用和注释，而智能体则负责在这些工具中进行选择、配置、解读输出结果并决定后续步骤。这使得语言模型的推理仅限于可靠的判断层，并保留了现有工具所保证的可重复性。

Agent reasoning is powered by a consortium of fine-tuned, domain-specialized large language models coordinated by a central reasoning LLM, executing locally so no sequencing data leaves the operating environment, under human-in-the-loop orchestration.

智能体的推理能力由一组经过微调的领域专用大语言模型提供支持，并由一个中央推理大语言模型进行协调。所有操作均在本地执行，确保测序数据不会离开操作环境，并在“人在回路”（human-in-the-loop）的编排下运行。

Evaluation shows agent-generated configurations are concordant with expert practice, that an explicit filtering ledger renders inspectable what filtering otherwise removes without trace, and that cross-stage anomaly detection surfaces conditions execution monitoring misses. BaseCamp offers a generalizable blueprint for agentic automation of scientific data pipelines.

评估结果表明，智能体生成的配置与专家实践一致；显式的过滤账本使得原本会被无痕移除的过滤内容变得可审查；跨阶段异常检测能够发现常规执行监控所遗漏的问题。BaseCamp 为科学数据流水线的智能体自动化提供了一个可推广的蓝图。