在本报告中，开发了两种算法，算法1和算法2。算法1是为了寻找具有短活性区域的全长蛋白质的多肽链的相互作用而开发的。算法2用于确定全长蛋白在N端到C端方向形成二聚体时相互作用最活跃的位点。采用Mdm2、Nap1、P53蛋白进行数值计算。对现代蛋白质组学来说，蛋白质相互作用的研究和预测是非常重要的任务，因为它们决定了蛋白质从细胞到整个生物水平的功能。对于结构已知的蛋白质，根据其三级结构构象的已知数据寻找分子间的相互作用，归结为寻找相互作用的两个分子表面截面的几何互补性，并对其接触进行建模，即所谓的分子对接。分子对接的任务是构象搜索算法的任务，由于蛋白质分子扭转角度的变化，该算法简化为对所形成的生物复合物构象空间的搜索。现代构象搜索算法在大多数情况下发现的构象通常在相对较短的时间内接近实验发现的结构。然而，也有一些因素对对接的成功有重大影响，这些因素在标准算法中往往没有考虑到。其中一个因素是目标蛋白的构象迁移。 The mobility range can be different beginning with a small adjustment of the side chains and ending with scale domain movements. These movements play an important role. At first glance, the most logical solution to this problem is to take into account the mobility of the protein in a docking program. Unfortunately, modern computational tools do not allow such modeling to be performed in an acceptable time frame since a protein molecule is very large, and allowing for mobility over all degrees of freedom can lead to a so-called combinatorial explosion (an astronomical increase in the number of possible variants). Only in some programs is there a limited mobility of protein binding sites (usually at the level of a small adaptation of conformations of the side chains of the active center residues). Another approach to this problem consists in docking the same protein in several different conformations and then selecting the best solutions from each docking run. The third approach is to find a universal structure of the target protein in which docking would produce fairly good results for different classes of ligands. In this case, the number of missed (but correct) solutions decreases, but the number of incorrect options also increases significantly. It should also be noted that most programs for the theoretical docking of proteins work according to the following principle: one protein is fixed in space, and the second is rotated around it in a variety of ways. At the same time, for each rotation configuration, estimates are made for the evaluation function. The evaluation function is based on surface complementarity (the mutual correspondence of complementary structures (macromolecules, radicals), determined by their chemical properties), electrostatic interactions, van der Waals repulsion and so on. The problem with this approach is that calculations throughout the configuration space require a lot of time, rarely leading to a single solution, which in turn does not allow us to speak of the uniqueness of the target protein and ligand interaction variant. So in the work while modeling by the methods of molecular dynamics, from 200 to 10 000 possible combinations of the formation of a protein complex with a ligand were found. Such a large number of modifications, along with the lack of a criterion for selecting the most probable variants of the bound structures of biological complexes (which would allow a radical reduction in their number) makes it very difficult to interpret the theoretical results obtained for practical use, namely, the finding of catalytic centers and a qualitative assessment of the dissociation constant of interacting substances. In contrast to the above computer simulation algorithms, mathematical algorithms have been developed in this chapter that allow determining the detection of proteins active regions and detecting the stability of different regions of protein complexes (linear docking) by analyzing the potential energy matrix of pairwise electrostatic interaction between different sites of the biological complex, such as the homodimer of the histone chaperone Nap1-Nap1, the heterodimer of the p53 Mdm2 proteins, and the homodimer Mdm2 Mdm2, which are responsible for the entry of a whole protein molecule into biochemical reactions.
T V Koshlan.