bioRxiv💻 bioinformatics

Dingent: An Easily Deployable Database Retrieval and Integration Agent framework

This paper introduces Dingent, a novel and easily deployable framework that bridges the gap in existing solutions by providing a configurable, web-based agent system for natural language data retrieval and integration across diverse sources, demonstrating its versatility through successful applications in multiple scenarios and potential for fields like earth sciences.

Kong, D., Bei, S., Wu, Y., Tang, B., Zhao, W.

Published 2026-03-20
📖 5 min read🧠 Deep dive
View on bioRxiv ↗PDF ↗
⚕️

This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

Imagine you are a researcher trying to find a specific piece of information. In the past, you might have had to visit five different libraries, learn five different cataloging systems, and ask five different librarians in five different languages just to get one answer.

Dingent is like a super-smart, multi-lingual concierge who solves this problem. It is a new software framework that lets scientists build their own "AI assistants" without needing to be computer programmers.

Here is a simple breakdown of how it works, using everyday analogies:

1. The Problem: The "Library of Babel"

Right now, there are thousands of biological databases (like giant digital libraries for DNA, proteins, and diseases). Some are organized like a spreadsheet (MySQL), some like a search engine (Elastic Search), and others are custom-built.

  • The Old Way: To ask a question, you needed a programmer to write complex code to connect your question to the right database. If you wanted to ask a question about dogs and genes, you needed two different programs talking to each other.
  • The Gap: Existing tools were either too hard for non-coders to use or too rigid to handle different types of data.

2. The Solution: Dingent (The "Lego Kit" for AI)

Dingent is a configurable toolkit. Think of it not as a finished product, but as a box of high-tech Legos and a set of instructions.

  • No Coding Required: Instead of writing code, a scientist uses a visual "dashboard" (like a drag-and-drop menu) to connect the dots.
  • The "One-Stop" Shop: You can plug in a database about dogs, a database about human genes, and a tool for analyzing protein structures. Dingent connects them all into one friendly chat window.

3. How It Works: The "Smart Kitchen" Analogy

Imagine you walk into a restaurant and order a complex dish: "I want a soup made from ingredients found in the ocean, but I want it to taste like a forest."

  • The Chef (The AI Agent): Dingent is the head chef. You speak to it in plain English.
  • The Pantry (The Plugins): The kitchen has different stations. One station handles seafood (Database A), another handles mushrooms (Database B), and another handles spices (Custom Tools).
  • The Recipe (The Workflow): You don't tell the chef how to chop the onions. You just say what you want. Dingent's "brain" (the execution engine) figures out the recipe:
    1. Go to the seafood station and get the fish.
    2. Go to the mushroom station and get the fungi.
    3. Combine them and cook.
    4. Serve the result to you.

If you ask, "What dog breeds need daily grooming?", Dingent knows to send that question to the "Dog Database" station. If you ask, "Find the gene TP53," it knows to send that to the "Biomarker Database" station. It routes your question to the right expert automatically.

4. Real-World Examples from the Paper

The authors tested this "Concierge" in three ways:

  • Scenario A: The Specialist (Single Database)
    They built a bot just for GenBase (a nucleotide database). You can ask, "Show me the DNA sequences for Ciona savignyi," and the bot instantly pulls up 88,000 results in a neat table. It's like having a dedicated librarian who only knows that one bookshelf but knows it perfectly.

  • Scenario B: The Traffic Controller (Multiple Databases)
    They connected three different databases: one for biomarkers, one for dog traits, and one for genetic sequences.

    • If you ask about dogs, the bot routes you to the dog database.
    • If you ask about biomarkers, it routes you to the biomarker database.
    • It acts like a smart switchboard operator, ensuring you talk to the right person.

  • Scenario C: The Detective (Connecting the Dots)
    This is the most powerful part. They asked a two-step question: "What causes degenerative myelopathy in dogs, and are those genes also biomarkers?"

    1. The bot first checks the Dog Database and finds the genes SOD1 and SP110.
    2. It then takes those names and automatically checks the Biomarker Database.
    3. It reports back: "Yes, SOD1 is a known biomarker for this disease."
    • The Magic: The bot connected two different libraries to solve a puzzle that would have taken a human hours to cross-reference.

5. Why This Matters

  • Democratization: You don't need to be a software engineer to build powerful AI tools. A biologist can build their own custom search engine in minutes.
  • Speed: It uses "caching" (like remembering a recipe you've made before) so it doesn't waste time rebuilding the same connections.
  • Future-Proof: If a new database comes out tomorrow, you just plug it into the system like adding a new appliance to your kitchen.

The Bottom Line

Dingent turns the chaotic, confusing world of scattered scientific data into a single, friendly conversation. It allows researchers to stop wrestling with code and start asking questions, letting the AI do the heavy lifting of finding, connecting, and summarizing the answers.

Note: The paper mentions that while it's great at finding data, it's still learning how to do complex data analysis and handle multiple users securely, but it's a huge step forward for making science accessible.

Drowning in papers in your field?

Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.

Try Digest →

More like this

Functional-space alignment resolves the eco-evolutionary landscape of siderophore biosynthesis across bacteria

By introducing a functional-space alignment framework (BGC Block Aligner) and a curated benchmark (SideroBank), this study reveals that siderophore biosynthesis is a pervasive trait across over 60% of bacterial genomes, where ecological lifestyle rather than phylogenetic relatedness drives the global distribution and evolution of these iron-acquisition systems.

Shao, J., Wu, Y., Tian, S. + 11 more2026-04-15💻 bioinformatics

Exploring molecular signatures of senescence with markeR, an R toolkit for evaluating gene sets as phenotypic markers

The authors developed markeR, an open-source R toolkit for systematically evaluating and benchmarking gene sets as phenotypic markers across diverse datasets, demonstrating its utility through a comprehensive analysis of senescence signatures that revealed significant variability in their performance and highlighted the tool's capacity to uncover context-specific biological insights.

Martins-Silva, R., Kaizeler, A., Barbosa-Morais, N. L.2026-04-15💻 bioinformatics

TFBindFormer: A Cross-Attention Transformer for Transcription Factor-DNA Binding Prediction

TFBindFormer is a hybrid cross-attention transformer that significantly improves the accuracy and scalability of transcription factor-DNA binding predictions by explicitly integrating DNA genomic features with TF-specific protein sequence and structural information, outperforming existing DNA-only models across diverse cell types and genomic contexts.

Liu, P., Wang, L., Basnet, S. + 1 more2026-04-15💻 bioinformatics