How AI Creates New Biomaterials
AI replaces costly trial and error in the development of new biomaterials
DAN*NA, a Barcelona-based green bioengineering company, has garnered international recognition for its high-tech biomaterials. Its flagship product is a bio-PLA that is more flexible than market competitors while maintaining transparency. She has also patented a new biobased material for tissue regeneration and bioprinting.
DAN*NA’s reputation is not just the product of human ingenuity. Its AI material design software informs every step of its development process. By simulating a virtual “clone” of target materials, developers can predict the physical properties of new formulations without arduous experimental testing. DAN*NA’s system and others like it help developers identify how bio-based materials should be handled to match petrochemical materials in performance and properties.
The company’s computational approach to biomaterials development has propelled it to the forefront of biomaterials innovation. Over the past three years, it has acted as a leading partner in several major private and public projects, both in Spain and abroad. One is the European Union’s $6.9 million Catco2nvers Horizon 2020 research and innovation program, which creates new value-added chemicals using carbon emissions captured by bio- industry. DAN*NA is also involved in BIOCON-CO2, another project developing carbon conversion methods for the manufacture of biomaterials.
Machine learning in materials design isn’t new, but its commercial adoption is widening. Global patent applications for AI-based materials science have grown from around 1100 in 2011 to over 3000 in 2017. Although AI technology for innovation and discovery is already widely used in pharmaceutical products, the technique is migrating towards the design of industrial biomaterials.
What is the place of AI in research on biomaterials?
The role that AI could play in the transition to sustainability has been much discussed. Some believe that technology could guide the development of environmental policies. Others say it could be deployed as a computational aid for resource optimization or to identify more efficient industrial processes. One of the most important ways that AI software can support economy-wide decarbonization is by generating new sustainable and economically viable innovations in materials design.
Many product applications have strict requirements on the physical properties their materials must provide. In food packaging, plastic containers can keep out moisture and heat. Building materials must not corrode, crack or deform under weight. Despite the enormous technical advances made over the past two decades in the design of biomaterials, some renewable materials are still not perfect functional substitutes for legacy petrochemical derivatives. Since biomaterials are typically more expensive to manufacture than their petrochemical counterparts, matching (and ideally exceeding) the performance of legacy materials is key to capturing a larger market.
The physical properties of any material are based on their unique chemical compositions and molecular structures. Sometimes the way the molecular or nanoscale structure can matter more in determining mechanical behavior than the elements that make them up.
However, there are a dizzying number of possible molecular combinations. This motivates researchers to adopt AI to generate promising configurations.
IBM, AI and sustainable computer chips
IBM is at the forefront of AI design for durable components in microchips. “Photoacid generators” (PAGs) are a photosensitive material used in the manufacture of computer chips. The chemical formulas of current PAGs are toxic to the environment, so the computer industry is looking for greener alternatives.
PAGs allow tiny chips to be engraved with maximum precision. Ultraviolet light is projected through a mask cut in a wiring pattern. When the light pattern hits a PAG-coated chip, the PAG exposed to the light breaks down into acids that eat away at the base material of the chip, leaving behind a perfect pattern.
The process of developing greener PAGs is expensive, especially if guided only by trial, error and human intuition. Instead, researchers are using AI to find and combine sustainable materials into a new type of PAG that is both made from renewable materials and has all the right optical properties. The team used IBM’s Deep Search AI to search scientific papers for candidate organic materials. Then, researchers input them into IBM’s “Intelligent Simulation” generative AI. Using these, the software offers suggestions on how these materials might be structured at the molecular level to achieve properties that make them effective PAGs. Another IBM technology then explores which of these outputs might work best.
BioBTX and AI for biochemical descaling
IBM also uses machine learning to choose the best methods for manufacturing PAGs. Another company that has used AI to find optimal methods of producing biomaterials is German startup BioBTX.
BioTX wanted to find cost-effective production methods for bio-based flavor ingredients. Aromatics are a group of chemicals used in clothing, pharmaceuticals, cosmetics, computers, wind turbines, paints, vehicle components, and sports equipment. Currently, aromatics are made from petrochemical derivatives, but BioBTX has been researching circular and bio-based renewable versions. He found a way to break down a waste product from biodiesel production, glycerin, into three key aromatic chemical inputs: benzene, toluene and xylene.
Bio-based versions of benzene, toluene and xylene must be functionally indistinguishable from their petrochemical versions. However, any manufacturing method must also be cost effective. The problem was that there are about 5 million potential methods to convert glycerin into three flavor ingredients. Each uses different raw materials, types of catalysts and temperatures. It is not obvious which one would be economically viable and physically optimal without trying each one. Testing even a fraction of the possibilities in the laboratory would be prohibitively expensive.
To circumvent the problem, BioBTX has teamed up with the University of Groningen on an algorithm that simulates real experiments on glycerine degradation methods and predicts their outcome. Based on the suggestions, the company set up a pilot plant to produce its first batches of bio-based flavor compounds. Now are working on the construction of a full-scale plant which is expected to come into operation by 2023.
The problem of chemical combinations also plagues companies working at the intersection of green chemistry and synthetic biology. Arzeda, a University of Washington synbio spin-off created in 2008, has created a protein design platform to predict the properties of fermented chemicals. What is unique about their software is that it was able to generate entirely new molecules with properties not found in any existing synthetic or natural material. The company takes these custom-designed molecules and produces them at scale through bioreactor fermentation.
California-based Zymergen, which went public in April 2021, also combines machine learning with bioengineering and biomanufacturing. Using machine learning, they can learn how high-value chemicals grown inside the bodies of microbes will behave in a real-world application. This is useful for selecting strains to consider in the scaling step. Zymergen used this process to manufacture its bio-based electronic films.
Zymergen also uses its software to identify how genetic strains of microbes are associated with better biomanufacturing outcomes, such as higher yields. Computer simulations can also give an accurate picture of the trade-offs that may be associated with optimizing a given microbial trait. For example, a small DNA alteration may produce a microbe that needs less sugar input but may take longer to produce the target chemical.
AI-based bio-design is at the heart of Zymergen’s business model. The company sells its simulation services to its customers in addition to using them to develop and manufacture new products in-house.
Industry and academia
Computational chemistry and industrial applications of machine learning are emerging fields. The biomaterials sector is also in its infancy. As a result, industry-university collaborations are being forged to refine software and expand databases of materials to train them.
In 2017, Toyota Research Institute (TRI) and Northwestern University embarked on the Accelerated Materials Design and Discovery (AMDD) project to apply AI to advanced materials research. The program researched new materials for a low-carbon automotive industry. In the first four years, he successfully predicted 19 entirely new materials using software developed by Soicheia Inc. The project built on Soicheia’s gargantuan database of over 200 million nanomaterials . The researchers generated nanoparticles of different compositions, structures, sizes and shapes by feeding the data into a machine learning algorithm with basic physical and chemical principles. To date, he has published over 150 academic papers on low carbon battery and fuel cell materials.
The Toyota Research Institute project received an initial investment that year of $35 million over four years. The project received a new funding injection of $36 million in 2021 to continue its research with several universities around the world, including the California Institute of Technology, Carnegie Mellon and MIT. Professor Yang Shao-Horn, who is involved in the research, commented: “We are working with TRI to bring together polymer synthesis, rapid robotic experiments, molecular simulation and AI to establish new design rules. for polymers.
Machine learning offers a path to increased efficiency in the fundamental research of innovative materials. This is critical for the biomaterials industry seeking to overcome market perceptions of substandard biobased products. However, the uses of machine learning go beyond the rapid prediction of physical performance and economic viability of molecular combinations. This could give the bio-based sector an edge over the synthetic chemicals industry by generating entirely new materials with unique properties that are difficult to derive from petrochemicals.