How to Use NOS Online
Try NOS
Commercially Available Online Web Server
NOS: A Breakthrough in Protein Design
NOS (diffusioN Optimized Sampling), a new gradient-guidance method for protein design that works directly in sequence space using discrete diffusion models. This approach represents a significant shift from traditional methods that rely on protein structure, circumventing major limitations such as the scarcity of high-quality structural data and the challenges of inverse design. NOS is a core component of a broader procedure called LaMBO-2, which enables multi-objective optimization and edit-based constraints for practical protein engineering tasks.
In exploratory in vitro experiments, LaMBO-2 achieved a 99% expression rate and a 40% binding rate for antibodies designed to target several therapeutic antigens. This demonstrates the power of a sequence-based approach to generate highly enriched libraries of functional protein candidates.
How NOS Works
NOS operates as an iterative refinement process on discrete protein sequences. Instead of working with a continuous representation of a protein's structure, NOS directly manipulates the sequence by following gradients in the hidden states of its denoising network. This is different from autoregressive models, which only make local changes, as NOS can refine the entire sequence at once.
Gradient Guidance: At each step of the diffusion process, NOS uses gradients to modify the protein sequence, guiding it toward a desired function while simultaneously maintaining the sequence's overall plausibility.
Saliency Maps for Targeted Editing: The method uses saliency maps to determine which amino acid positions are most important to edit to improve a specific objective. This is a crucial feature for constrained design tasks, as it allows researchers to focus a limited "edit budget" on the most impactful positions, such as the Complementarity-Determining Regions (CDRs) of an antibody.
LaMBO-2 Integration: By combining NOS with a multi-objective Bayesian optimization procedure (LaMBO), the method can efficiently search for new sequences that offer an optimal balance between competing goals, such as high binding affinity and high expression yield.
What is Tamarind Bio?
Tamarind Bio is a pioneering no-code bioinformatics platform built to democratize access to powerful computational tools for life scientists and researchers. Recognizing that many cutting-edge machine learning models are often difficult to deploy and use, Tamarind provides an intuitive, web-based environment that completely abstracts away the complexities of high-performance computing, software dependencies, and command-line interfaces.
The platform is designed provide easy access to biologists, chemists, and other researchers who may not have a background in programming or cloud infrastructure but want to run experimental models with their data. Key features include a user-friendly graphical interface for setting up and launching experiments, a robust API for integration into existing research pipelines, and an automated system for managing and scaling computational resources. By handling the technical heavy lifting, Tamarind empowers researchers to concentrate on their scientific questions and accelerate the pace of discovery.
Accelerating Discovery with NOS on Tamarind Bio
Using the NOS framework on a platform like Tamarind could accelerate the discovery and optimization of protein therapeutics in several ways:
Efficient Lead Optimization: Researchers can use NOS to generate and test new antibody variants that are optimized for multiple objectives, such as expression yield and binding affinity. By leveraging the platform's computational power, they can quickly explore a vast sequence space and identify promising candidates for experimental validation.
Intuitive Design with Constraints: The use of saliency maps makes the design process more intuitive, allowing researchers to see which parts of a sequence are most critical to a given objective. This insight, combined with the ability to impose constraints like an "edit budget," enables a highly controlled and efficient design workflow that is crucial for antibody lead optimization.
High-Throughput and Automation: With Tamarind, researchers can automate the entire NOS workflow, from initial seed selection to the generation of a refined library of designs. The platform handles the computational heavy lifting, enabling scientists to produce high-quality protein libraries in silico and significantly reduce the time and cost associated with traditional wet-lab methods.
How to Use NOS on Tamarind Bio
To leverage NOS's power, a researcher could follow this streamlined workflow:
Access the Platform: Begin by logging in to the tamarind.bio website.
Select NOS: From the list of available computational models, choose the NOS tool.
Input a Seed Sequence and Objective: Start by providing an initial protein sequence (the seed) and a clear objective for optimization, such as improving binding affinity to a specific target.
Define Constraints: Specify any constraints for the design, including an edit budget (the number of allowed amino acid changes) and any sequence liabilities to avoid.
Run LaMBO-2: The platform would run the LaMBO-2 framework with NOS. This process uses saliency maps to automatically select the most impactful positions for editing and then applies discrete diffusion with gradient guidance to generate new, optimized sequences.
Evaluate and Select Candidates: The platform would return a ranked library of new sequences that are optimized for the specified objectives while adhering to the defined constraints. These candidates are highly enriched for functionality and can be directly used for experimental validation, bypassing the need for extensive high-throughput screening.