Lectin-Glycan Complexes: A Comprehensive Analysis of Docking Calculations

Lectins are a type of glycan-binding protein that noncovalently bind glycans. Carbohydrates are molecules consisting of sugar units joined together. Glycans are carbohydrates. Hence, glycans are also sugars. Lectins and lectin-glycan complexes have a range of biological roles and can be found in animals (including humans), plants, bacteria, viruses, and yeasts and fungi.1 Many scientists focus on the computational study of these complexes due to their intricate roles in many living organisms. Computational study is important in furthering our knowledge of lectin-glycan complexes and other such protein complexes. However, computational study is not perfect. There are many challenges in computational study, especially in the docking of ligands to receptors. In glycans, there are usually a high number of hydroxyl (OH) groups that affect docking; there could be surrounding ions; the rings in glycans can have CH-π stacking interactions with aromatic residues and cause issues; and glycans larger than one subunit have bonds between subunits that allow them to twist. These represent just a few challenges in docking. It is difficult for the software to accurately dock glycans to corresponding receptors because of these challenges. So, the purpose of this study was to try to evaluate the performance of the docking program Autodock Vina (referred as Vina) and Vina-carb on a large dataset of docking problems and propose a workflow for effective docking of glycan ligands.2 We have looked into the effect of glycan size, seed value (a random starting point for docking calculations), and Carbohydrate Intrinsic (CHI) energy functions in glycosidic linkages.3 We tried using CHI values in Vina-Carb that mimicked Autodock Vina. We saw that in almost all cases Vina-Carb did better, even if it was a marginal difference. Then we tried optimizing the CHI values CHI coefficient and CHI cutoff.3 We did see some patterns emerge for specific values. We also used a random seed for calculations but did not see much of a difference in using a random seed for calculations. There were some improvements and surprises. Overall, we know that optimizing docking software is a challenge, but doing so will improve research for many scientists. More calculations will be done in the future because they will be worthwhile. We originally sought to analyze Vina-Carb to make it better. Such research will help improve future computational study. We have already seen some parameters that have promise for further investigation