De novo design is a scientific approach that involves creating something entirely new, “from scratch,” rather than modifying existing structures. This method applies across various scientific disciplines, including structural bioinformatics, drug discovery, and materials science. It represents a shift from traditional discovery methods, enabling researchers to envision and build novel systems with specific, desired properties. This strategy transforms how scientists approach complex problems, opening new avenues for discovery and development.
Building from Scratch
De novo design constructs molecules or systems with targeted properties from their fundamental components, rather than adapting pre-existing ones. Traditional scientific approaches often involve screening large libraries of existing compounds or making small, iterative changes to known structures through trial and error. While these methods have yielded successes, they are often limited by the available chemical space or by the inherent properties of the starting materials.
De novo design, in contrast, bypasses these limitations by allowing for the creation of entirely new chemical entities. This “building from scratch” philosophy enables scientists to explore a much broader range of possibilities, leading to novel structures that might not be found in nature or existing databases.
By focusing on designing in desired attributes from the ground up, this approach offers increased efficiency and improved success rates in developing molecules with specific functions, such as binding affinity to a target protein. It allows for the systematic construction of systems tailored to precise requirements, which can lead to higher specificity and efficacy in various applications.
The Design Process
The de novo design process is typically iterative and heavily relies on computational tools to generate novel structures based on predefined criteria. The first step involves defining the specific problem and the desired properties of the molecule or system to be designed. This includes outlining constraints such as a predefined solubility range, toxicity thresholds, or the inclusion of specific chemical groups in the structure.
Once the design problem is clearly articulated, computational algorithms and simulations are employed to generate potential molecular structures. These tools, often incorporating artificial intelligence and machine learning techniques, explore a vast chemical space to propose new compounds. Methods like atom-based, fragment-based, or reaction-based de novo design are utilized to assemble novel ligands by combining molecular fragments or by simulating chemical reactions. For instance, fragment-based methods might involve growing, linking, or merging binding fragments within a target’s active site.
After generating a pool of candidate structures, these are rigorously evaluated using scoring functions that predict their binding affinity, drug-likeness, and other pharmacokinetic properties. This computational evaluation helps narrow down the vast number of possibilities to the most promising candidates. Machine learning models, including recurrent neural networks and generative adversarial networks, are increasingly used to predict properties like affinity or toxicity, further refining the selection process.
The most promising computationally designed candidates are then synthesized in the laboratory. This experimental synthesis can be a complex undertaking, as the computational methods do not always guarantee synthetic accessibility. Following synthesis, the designed molecules undergo experimental validation to confirm their predicted properties and biological activities. This iterative cycle of computational design, synthesis, and experimental testing allows for continuous refinement and optimization, ensuring that the final molecules meet the desired specifications and exhibit the intended functionality.
Transforming Science and Medicine
De novo design is significantly impacting various scientific and medical fields by enabling the creation of tailored solutions. In drug discovery, this approach allows for the generation of entirely new chemical entities optimized for specific biological targets, rather than relying on existing compounds. For example, de novo drug design has been instrumental in developing novel drug molecules for newly discovered diseases, such as COVID-19, accelerating the development of treatments. This leads to drugs with specific therapeutic effects and potentially fewer side effects, as they are designed to interact precisely with their intended targets.
In protein engineering, de novo design involves creating proteins with sequences unrelated to those found in nature, based on fundamental biophysical principles. This allows for the development of novel proteins with enhanced stability or catalytic activity for industrial or medical uses. For instance, scientists have designed small, hyperstable proteins that bind to the SARS-CoV-2 spike protein with high affinity, demonstrating their potential as diagnostics, therapeutics, and vaccine components. These designed proteins can be produced efficiently in bacteria like E. coli, offering a scalable alternative to traditional antibody-based treatments.
The field of materials science also benefits from de novo design, leading to the development of new materials with customized properties. This includes creating materials that are stronger, lighter, more conductive, or capable of specific functions like gas storage. For example, de novo design has been used to identify highly porous solids that can store over 150 times their own volume of methane, addressing challenges in sustainable energy and fuel storage. Furthermore, this approach is being applied to engineer living materials with tunable mechanical properties and programmable functionalities, such as 3D printability or binding to nanoparticles, by discovering and designing protein nanofibers.