Generative AI and Proteins

 Always an interest of mine, often mentioned here, approach should be interesting

Generative AI is dreaming up new proteins      by Laura Howes  in Cen.acs.org

April 10, 2023 | A version of this story appeared in Volume 101, Issue 12

In the past year, phrases like learned language models, diffusion, and hallucination have gained new meanings in popular culture as artificial intelligence has started taking over mundane tasks. Today, users can log on to an AI-powered chatbot and ask it to draft texts based on simple prompts. They can then use text-to-image services to create illustrations and new images to accompany the dreamed-up words.

But beyond these consumer applications, algorithmic approaches are helping researchers create a whole world of new proteins—proteins that could become vaccines, biologic therapies, materials, or tools for bioremediation.

A few years ago, C&EN chatted with David Baker of the University of Washington about a host of topics, including de novo protein design, which is designing new proteins from scratch rather than adjusting existing ones. Back then, he said he tried not to look too far into the future. Too much could change; too much was uncertain. That has never been truer.

De novo protein design has reached an inflection point, researchers say. AI-powered protein design is becoming very real and very usable, thanks to technological advances in the development of algorithms and the hardware that runs them.

Protein science itself was uniquely positioned to take advantage of these advances because of the enormous amounts of work carried out over the past 50 years to curate and annotate biological data.

“Every time there is a new method in computer vision or natural language processing, we are in a race to try to transfer it to biology,” says protein designer Noelia Ferruz at the Institute of Molecular Biology of Barcelona. “I guess it’s the perfect moment, because we’re seeing an AI revolution in every field.”

THE PROTEIN DESIGN PUZZLE

Proteins are hugely variable and incredibly specialized. They can form large, complicated complexes that mediate biological functions, or they can exist as small peptides that merely send signals from place to place. Proteins move and interact. They bind to and modify one another. They form part of the complex molecular dance we call life.

But proteins are also just molecules. They are composed of amino acid building blocks stuck together using amide linkages to create polymer chains. These polymers can curl up to build various shapes, chemical environments, forms, and functions, depending on where they sit and the interactions between the side chains in the molecules. For example, a hydrophobic portion might be found curled up inside the protein it’s part of or buried in a fatty membrane.  ... ' 

Large language models also Work for Protein Structures

We asked this question long ago, usefully answered?  Very big deal if so.

THE LANGUAGE OF BIOCHEMISTRY —

Large language models also work for protein structures

Training on raw protein sequences allows the AI to make inferences about structure.

JOHN TIMMER - 3/16/2023, 3:01 PM

The success of ChatGPT and its competitors is based on what's termed emergent behaviors. These systems, called large language models (LLMs), weren't trained to output natural-sounding language (or effective malware); they were simply tasked with tracking the statistics of word usage. But, given a large enough training set of language samples and a sufficiently complex neural network, their training resulted in an internal representation that "understood" English usage and a large compendium of facts. Their complex behavior emerged from a far simpler training.

A team at Meta has now reasoned that this sort of emergent understanding shouldn't be limited to languages. So it has trained an LLM on the statistics of the appearance of amino acids within proteins and used the system's internal representation of what it learned to extract information about the structure of those proteins. The result is not quite as good as the best competing AI systems for predicting protein structures, but it's considerably faster and still getting better.  

LLMs: Not just for language

The first thing you need to know to understand this work is that, while the term "language" in the name "LLM" refers to their original development for language processing tasks, they can potentially be used for a variety of purposes. So, while language processing is a common use case for LLMs, these models have other capabilities as well. In fact, the term "Large" is far more informative, in that all LLMs have a large number of nodes—the "neurons" in a neural network—and an even larger number of values that describe the weights of the connections among those nodes. While they were first developed to process language, they can potentially be used for a variety of tasks. .... ' 

Computer Models Determine Drug Candidate's Ability to Bind to Proteins

Pharma applications? 

 Computer Models Determine Drug Candidate's Ability to Bind to Proteins   By University of Arkansas, January 17, 2023

The research focuses on computational simulations of diseases, including coronavirus.

University of Arkansas (U of A) researchers have developed computer models for calculating a drug candidate's protein-binding affinity.

Said U of A's Mahmoud Moradi, the method "assigns an effective energy to the ligand at every grid point in a coordinate system, which has its origin at the most likely location of the ligand when it is in its bound state."

The researchers produced a computationally efficient binding estimator using biased simulations and non-parametric re-weighting techniques, then applied orientation quaternion formalism to further define the ligand's conformational shifts while binding to targeted proteins.

They used the method to estimate binding affinity between human fibroblast growth factor 1 and heparin hexasaccharide 5 medication.

From University of Arkansas   

AI Helps Generate Building Blocks for New Drugs

Generating New Drugs

AI Helps Biotech Labs Generate the Building Blocks for New Drugs

By Adrianna Nine on December 6, 2022 at 7:56 am Comments

Myotis daubentonii. (Credit: Generate Biomedicines)

Proteins are an essential part of life. Not only do they function as the “building blocks” for living organisms, but they also perform nearly every cellular task, from waste management to tissue repair. It tracks, then, that pharmaceuticals often contain or “target” proteins in an attempt to change or eliminate symptoms or disease within the body. There’s just one little problem: The only proteins we can use to create drugs are the ones we know.

But what if we could add new proteins to the resource pile? Two biotech labs have begun using artificial intelligence (AI) to generate novel proteins with the goal of incorporating them into medicine. Chroma, a program by Boston-based Generate Biomedicines, and RoseTTAFold Diffusion, a program out of the University of Washington’s Baker Lab, are often compared with trendy AI art generators thanks to their ability to create new concepts seemingly out of thin air. In fact, Chroma has even been referred to as “the DALL-E 2 of biology.”

AI protein generators, a type of denoising diffusion probabilistic model (DDPM), are trained to pull samples from complex datasets. They then use those samples—plus added noise—to construct a product that matches a given prompt. Although art generators are the best-known DDPMs, AI isn’t only good for producing images on demand; they’ve also been found to be an ideal method of generating diverse protein models. In a preprint on bioRxiv, Baker Labs scientists describe how RoseTTAFold can devise protein structures with virtually any desired size, shape, or function, as well as with any required constraints. ... ' 

Protein Behavior

 Multiscale Model of Protein Behavior Linked to Cancer-Causing Mutations

Lawrence Livermore National Laboratory, January 10, 2022

Researchers at the U.S. Department of Energy's Lawrence Livermore National Laboratory (LLNL) worked with a multi-institutional team on a multiscale model of the behavior of RAS proteins associated with cancer-inducing mutations. The researchers said the Multiscale Machine-Learned Modeling Infrastructure (MuMMI) could inform experiments and enhance understanding of RAS protein binding. MuMMI can extract insights at two temporal and spatial scales, allowing researchers to analyze thousands of RAS-lipid compositions and observe interaction patterns and RAS orientations. The team used LLNL's Sierra supercomputer to model a one-micron-by-one micron lipid patch to visualize the interaction of RAS proteins with lipids. The researchers said the experiments' findings will feed back into the MuMMI model, boosting its accuracy through a validation loop.  .... '

Tool Predicts Where Coronavirus Binds to Human Proteins

Computational tool for predicting binding sites.

Tool Predicts Where Coronavirus Binds to Human Proteins

Cornell Chronicle, Krishna Ramanujan, November 29, 2021

Cornell University researchers have developed a computational tool for predicting binding sites on the surfaces of human and COVID-19 viral proteins. A user-friendly Web server also provided by the tool’s developers shows all protein structures, so virologists and clinicians can determine whether current drugs, or those under development, will bind to them. "The tool we developed to predict protein-to-protein interfaces is the most accurate, and we can use it to make the most informed predictions for any interactions," said Cornell's Haiyuan Yu. Yu added that the tool also provides structural models for predicting how genetic mutations to proteins in individuals potentially impact viral interactions.

https://news.cornell.edu/stories/2021/11/new-tool-predicts-where-coronavirus-binds-human-proteins