The molecular structure of a drug, and each fragment that composes it plays its own role, so researchers envision combining or extending different structural fragments in order to obtain new drug molecules. This is theoretically feasible. According to this idea, Jencks et al. built the theoretical framework of FBDD in 1981, and as a result, the FBDD method began to be applied in the field of drug design. Different from screening millions of compound databases to directly search for drug-size molecules (drug-size, that is, molecules of the same size as the drug molecule), the FBDD method starts from screening small fragment structures, which are usually Contains less than 20 heavy atoms. The more complex the molecule becomes (the molecular structure becomes larger and the features become more), the more interactions with the protein are possible, and the interaction can be any protein contact, such as hydrogen bond interactions and hydrophobic interactions. It can be a favorable or unfavorable interaction. For a molecule that may become a potential drug, increasing beneficial interactions is the key. Although small fragments themselves have fewer protein interactions (weak binding), their advantage is that they can bind to multiple sites or multiple proteins of a protein. This way, screening with fragments as a starting point can increase the hit rate. If the hit fragments can be connected or combined into a large molecule of considerable complexity through a certain method, that is, reach the size of the drug, it is likely to become the potential drug of the target, and exert as many interactions with the target as possible.
Figure 1. Fragment-based Drug Design.(From Nature Reviews | Drug Discovery)
Compared with macromolecule screening, fragment screening has its practical advantages:
The theory of FBDD believes that the active pockets of many drug targets are composed of multiple subactive cavities. The fragments of the active compound obtained by HTS often cannot bind well to the subactive cavities of the target protein. The optimization of the product often affects the entire molecule, and even changes the binding position to the target, leading to loss of activity. The FBDD method connects the fragments that specifically bind to each subactive cavity of the target protein with suitable linkers to assemble them into compounds with high activity. Therefore, drugs designed by the FBDD method often have the characteristics of high activity and high selectivity.
Figure 2.FBDD pipeline.
The first step of the FBDD method is to establish a reasonable fragment library. The establishment of the fragment library can refer to some principles of the establishment of HTS compound library. Of course, the fragment library of FBDD also has its own characteristics: the molecular weight of the fragment is small, the number of hydrogen bond donors and acceptors is small, and the binding force to the target is weak. In addition, the fragments must have good solubility in order to be in high concentration solutions. Detect fragment activity in the library to meet the "three rules", with a molecular weight of less than 300, hydrogen bond donors and acceptors not exceeding 3, the number of rotatable bonds not exceeding 3, and cLogP less than 3. When the fragment library is established, the most critical step is to screen and identify active fragments that are weakly bound to the target protein. Currently, there are five major technologies for identifying fragments: biochemical detection, surface plasmon resonance technology (SPR), nuclear magnetic resonance technology (NMR), mass spectrometry (MS) technology, and X-ray single crystal diffraction (X-ray) technology. The fragments must be optimized and connected after identifying the active ones. The three main strategies are fragment-linking, fragment-merging and fragment-growing.
Artificial intelligence has attracted much attention in recent years and has successfully entered the field of drug discovery. Many machine learning methods, such as QSAR methods, SVMs or random forest methods, are established during the drug discovery process. New algorithms based on neural networks, such as deep neural networks, provide further improvements for attribute prediction, which have been revealed in many benchmark studies comparing deep learning and classic machine learning. The applicability of these new algorithms in many different applications has been proven, including physical and chemical properties, biological activity, and toxicity. In particular, because they benefit from increased computing power, they can simulate larger systems with more accurate methods.
|Project name||Fragment-Based Drug Design|
|Product delivery mode||The simulation results provide you with the raw data and analysis results of molecular dynamics.|
MedAI have cooperated with many universities or research institutions and can consult industry experts for their opinions and suggestions on some complex issues. Our drug design team is a highly professional, efficient and well-established R&D team. We will deliver tasks efficiently and on time, satisfying customers.
Tech Trends in Artificial Intelligence What's up, what's down, what lies ahead, only found in MedAI!