Graph Neural Networks as a Potential Tool in Improving Virtual Screening Programs.

Despite the increasing number of pharmaceutical companies, university laboratories and funding, less than one percent of initially researched drugs enter the commercial market. In this context, virtual screening (VS) has gained much attention due to several advantages, including timesaving, reduced reagent and consumable costs and the performance of selective analyses regarding the affinity between test molecules and pharmacological targets. Currently, VS is based mainly on algorithms that apply physical and chemistry principles and quantum mechanics to estimate molecule affinities and conformations, among others. Nevertheless, VS has not reached the expected results concerning the improvement of market-approved drugs, comprising less than twenty drugs that have reached this goal to date. In this context, graph neural networks (GNN), a recent deep-learning subtype, may comprise a powerful tool to improve VS results concerning natural products that may be used both simultaneously with standard algorithms or isolated. This review discusses the pros and cons of GNN applied to VS and the future perspectives of this learnable algorithm, which may revolutionize drug discovery if certain obstacles concerning spatial coordinates and adequate datasets, among others, can be overcome.

Quantification of schizophyllan directly from the fermented broth by ATR-FTIR and PLS regression.

Non-destructive methods that allow the quantification of bioproducts in a simple and quick manner during fermentation are extremely desirable from a practical point of view. Therefore, a 9 day fermentation experiment with Schizophyllum commune was carried out to investigate the possibility of using ATR-FTIR to quantify the schizophyllan biopolymer (SPG) directly from the culture medium. On each day, aliquots of the fermentation were taken, and the cell-free supernatant was analyzed by ATR-FTIR. The main objective of this step was to evaluate whether FTIR would be able to detect the appearance of specific peaks related to the production of SPG. The results of the PCA analysis showed that there was a reasonable separation of the days through the FTIR spectra. Then PCA-LDA was applied to the same dataset, which confirmed the formation of groups for each day of fermentation, after which, a calibration and test set was developed. Through a matrix generated by an experimental design with 2 factors and 5 levels, 25 samples were created with variations in the concentration of the culture medium and SPG. The ATR-FTIR spectra of this data set were modeled using PLS regression with backward selection of predictors. The results revealed that the amount of SPG produced can be quantified directly in the culture medium with excellent precision with R2CV = 0.951, R2P = 0.970, RMECV = 0.205 g, RMSEP = 0.170 g, RPDcv = 4.53 and RPDp = 5.88. The traditional method to quantify SPG is time consuming, requires several steps and uses solvents. In contrast, the method proposed in this work is a viable, faster, and a simpler alternative, which does not use reagents and does not require extensive pre-treatment of the samples.