Simplifying bioinformatics data analysis through conversation

Abstract

The burgeoning field of bioinformatics has been revolutionized by the rapid growth of omics data, providing insights into various biological processes. However, the complexity of bioinformatics tools and the rapidly evolving nature of data analysis pipelines present significant challenges for researchers, especially those lacking extensive programming expertise. To address these challenges, we introduce BioMANIA, an artificial intelligence-driven, natural language-oriented bioinformatics data analysis pipeline. BioMANIA comprises two key components: a ChatBot generation pipeline and a user-friendly ChatBot back-end service. The generation pipeline takes as input an open-source (e.g., hosted in GitHub) and well-documented (e.g., hosted in ReadTheDocs) Python tool, extracting API attributes and generating synthetic instructions that train a ChatBot to understand and perform specific data analysis tasks. We identified 11 common issues to provide a practical guideline for designing more ChatBot-compatible tools, which we categorized into five groups, while analyzing 12 well-documented open-source Python tools across various bioinformatics settings. The ChatBot service then assists users in selecting the appropriate analysis API and parameters, significantly reducing the programming barrier. We applied BioMANIA to analyze single-cell gene expression data, demonstrating its effectiveness in simplifying complex omics data analysis. BioMANIA has the potential to transform the landscape of bioinformatics research, making data analysis more accessible and accelerating discoveries in the field.1

Replicating a High-Impact Scientific Publication Using Systems of Large Language Models

“…Lastly, the core piece of our architecture: the analysis subsystem, had mixed performance results which varied depending on factors such as instruction complexity, degree of human involvement, and dataset size. As has been noted by others [13,27] code generation performance is meaningfully affected by instruction complexity. The first issue with making requests involving any level of complexity is that it results in increasingly inconsistent outputs.…”

“…The architecture we implored to answer this question is illustrated in Figure 1, of note is that numerous works exist [12][13][14] demonstrating the capabilities of LLMs to perform certain steps of this pipeline, and given the superior performance of GPT-4 across a range of benchmarks [15], we restricted ourselves to models built on Ope-nAI's platform. Specifically, we applied a system of text-generation and data analysis agents [14,15] to the task of replicating the methodology of a high impact journal article [16].…”

Leveraging large language models for data analysis automation

“…To this end, there are multiple proposed LMM-based solutions for data analysis that either benchmarked LLMs or built LMM-based solutions (12,13). In addition, there are specialized bioinformatics solutions relying on LLMs that bring prompt based solutions to enhance bioinformatics tasks (14)(15)(16)(17). Our main differentiating factor is building a more complete and ready use solution that integrates advanced prompt engineering solutions as well as interactivity.…”