We’re witnessing a fundamental shift in bio research, where artificial intelligence and engineering approaches are transforming traditional wet lab work into computational problems.[1] Yet this transformation has created a new challenge: the most powerful bio research tools, while more capable than ever, require substantial computational resources and complex technical setups that remain out of reach for many laboratories.
The reality of running state-of-the-art bioML models has become increasingly complex. Research teams need expensive GPU infrastructure just to get started, followed by intricate software deployments that demand significant technical expertise. The setup and maintenance alone can consume hours that could be better spent on actual research. As a result, many researchers spend more time wrestling with infrastructure than advancing their work. We’re changing that by collaborating with NoLabs on Lilypad's distributed compute network.
NoLabs is a comprehensive and highly scalable Open Source biolab platform that empowers researchers to run experiments with the latest state-of-the-art models, bioinformatics tools and scalable no-code workflow engine for bio research. For protein structure analysis, it offers powerful tools like ESMFold, RoseTTAFold and ProteinMPNN. The drug discovery pipeline incorporates DiffDock for protein-ligand docking, alongside BioBuddy AI Copilot for research assistance and REINVENT4 for molecular design. When it comes to molecular design, researchers can leverage RFdiffusion for structure generation and access specialized tools for small molecule design and optimization.
For example, researchers developing therapies for glioblastoma can use NoLabs to analyze mutations in the IDH1 enzyme, a common target in this aggressive brain cancer. They could utilize:
ESMFold to model the 3D structure of the mutated IDH1 protein.
REINVENT4 to rapidly generate novel small molecule inhibitors tailored to the modeled binding site.
Gene Ontology Analysis to predict the potential functional impact of these molecules on related biological pathways.
The NoLabs-Lilypad integration adds a crucial piece to scalable advanced bioML which is access to a flexible and on-demand GPU. This rapidly accelerates oncology research teams to identify and refine promising drug candidates, advancing the development of personalized treatments for glioblastoma.
Lilypad: Serverless, distributed on-demand compute network
Lilypad is pioneering a serverless, distributed compute network that is revolutionizing AI access by enabling internet-scale data processing for AI, ML, and other arbitrary computation. It empowers researchers with on-demand, cost-efficient access to high-performance computing and access of federated data securely, eliminating the need for upfront investments in infrastructure, coordination and setup overheads. With streamlined workflows, users can solely focus on discovery rather than logistical hurdles. Additionally, the capability for Model Monetisation provides a valuable proposition for Open source Developers to monetise their expertise and work while additionally providing provenance pipelines to data and IP rights to models.
Running on Lilypad’s Distributed Network
NoLabs integration with Lilypad's serverless GPU infrastructure makes powerful research tools instantly accessible. As a first step, we have converted individual bioML modules to Nolab-Lilypad Docker containers which have all the dependencies contained.
The workflow is straightforward: Researchers select the NoLab-Lilypad module, submit a job to Lilypad to request GPU, the deal is matched with a Resource provider and results are returned automatically. The impact on research teams is immediate and transformative. Gone are the days of software installation headaches and GPU requirements - researchers can now access these tools instantly without any configuration hassles.
Lilypad automatically handles resource allocation and enables parallel job execution, ensuring efficient utilization of computing power. From a cost perspective, teams only pay for the compute they actually use, eliminating the need for upfront infrastructure investments while allowing them to scale their costs based on actual usage.
See ESMFold in action below! 👇
This individual model execution of NoLabs and Lilypad modules is just a starting point.
We are now exploring how to scale and are testing wrapping the entire Nolabs application - not just individual modules - into a Lilypad Docker Container to ensure an even smoother and faster experience for researchers. Our aim is to support the NoLabs’ visual workflow which can connect multiple models into sophisticated pipelines on Lilypad. For instance, you can start with a protein structure prediction using ESMFold light, feed its output directly into DiffDock for ligand binding analysis, and visualize the entire protein-ligand complex - all through an intuitive drag-and-drop interface. This workflow capability transforms complex multi-step analyses into streamlined, automated processes which requires substantial GPU resources.
🧬
NoLabs workflow editor: visual orchestration of bioML modules
Looking Ahead: Scaling bioML in-silico experiments and creating economic incentive layers
The integration of NoLabs with Lilypad's distributed compute network lays the groundwork for transformative changes in how we conduct bio research. We envision small labs running hundreds of protein folding simulations simultaneously, university teams processing drug discovery pipelines without dedicated hardware investments, and biotech startups scaling their molecular design workflows instantly. Specifically for large and important in-silico experiments we can utilise these advantages to help researchers run experiments on scale.
Beyond expanding our network of compute providers and incorporating specialized bioML research tools, we’re building towards a future where AI tools and resources are accessible to every researcher, developer, and institution, while also creating IP, provenance and economic incentive layers - empowering the next wave of scientific discovery and technological progress.
Learn more about Lilypad
Ready to explore the world of distributed bio research with NoLabs? Join the Lilypad community and experience the power of decentralized computing for yourself.
Explore our docs
Connect with us on Discord
Together, let’s push the boundaries of scientific discovery and build a more open and accessible future for researchers.
About Lilypad
Lilypad is pioneering a serverless, distributed on-demand compute network that is revolutionizing AI access by enabling internet-scale data processing for AI, ML, and other arbitrary computation. Unleashing idle GPU processing power by leveraging decentralized infrastructure networks, Lilypad unlocks a new marketplace for compute, making AI more accessible, efficient, and transparent, enabling innovation without limits. Lilypad is empowering the intelligence era, solving the possible—one job at a time.
About NoLabs
NoLabs is pioneering the future of computational biology through its innovative open-source biolab platform. Founded by technical experts Timur Ishmuratov and Igor Bruev, the company is revolutionizing how researchers access and utilize advanced bioML tools. Through their partnership with Lilypad's distributed compute network, NoLabs is transforming the biotech research landscape by removing traditional infrastructure barriers and enabling researchers to focus on scientific discovery rather than technical setup.
Check out NoLabs GitHub
Join NoLabs Discord
Follow NoLabs on LinkedIn
References:
[1] The New Industrial Revolution: Bio x AI: https://a16z.com/the-new-industrial-revolution-bio-x-ai/
