Molecule docking has been regarded as a routine tool for drug discovery, but its accuracy highly depends on the reliability of scoring functions (SFs). With the rapid development of machine learning (ML) techniques, ML-based SFs have gradually emerged as a promising alternative for protein-ligand binding affinity prediction and virtual screening, and most of them have shown significantly better performance than a wide range of classical SFs. Emergence of more data-hungry deep learning (DL) approaches in recent years further fascinates the exploitation of more accurate SFs. Here, we summarize the progress of traditional ML-based SFs in the last few years and provide insights into recently developed DL-based SFs. We believe that the continuous improvement in ML-based SFs can surely guide the early-stage drug design and accelerate the discovery of new drugs. This article is categorized under:Computer and Information Science > Chemoinformaticsdeep learning, machine learning, molecular docking, scoring function, structure-based drug design | INTRODUCTIONTraditional drug discovery largely relies on the application of high-throughput screening, an experimental technique with acceptable performance but high cost and low efficiency. 1 With the rapid development of computational chemistry and computer technology, computer-aided drug design (CADD) has gradually emerged as a powerful technique in the design and development of new drug candidates in the past three decades. 2 Virtual screening (VS), an important branch of CADD, can enrich potential actives from large virtual compound libraries through in silico methods rather than real experiments, which can not only accelerate the process of drug discovery but also greatly reduce the time and resource cost. 3-5 Depending on whether the three-dimensional (3D) structure of a target is used or not, VS approaches can be classified into two major categories: ligand-based virtual screening (LBVS) and structure-based virtual screening (SBVS). 6 LBVS aims to discover active molecules through the models developed based on a set of known ligands of a target of interest, which may limit its capability to find novel chemotypes. Compared with LBVS, SBVS is considered to be a better choice to discover novel active compounds if the 3D structure of a given target is available. 7 Chao Shen and Junjie Ding are equivalent first authors.

This is the first study to extensively determine the effect of CYP3A4*1G and CYP3A5*3 genetic polymorphisms and hematocrit value on tacrolimus pharmacokinetics in Chinese renal transplant recipients. The findings suggest that CYP3A5*3 and CYP3A4*1G polymorphisms and hematocrit are determinant factors in the apparent clearance of tacrolimus. The initial dose design is mainly based on CYP3A5 and CYP3A4 genotypes as well as hematocrit. This result may also be useful for maintenance tacrolimus dose optimization and may help to avoid fluctuating tacrolimus levels and improve the efficacy and tolerability of tacrolimus in kidney transplant recipients.

WHAT IS ALREADY KNOWN ABOUT THIS SUBJECT• Mycophenolic acid (MPA) undergoes enterohepatic circulation (EHC) in the body and several population models have been proposed to describe this process using sparse data.• Recent studies in Whites have found that polymorphism in UGT1A9 could partly explain the large interindividual variability associated with the pharmacokinetics of MPA. WHAT THIS STUDY ADDS• A new population pharmacokinetic model for EHC combining MPA and its main glucuronide metabolite (MPAG) simultaneously was established based on physiological aspects of biliary excretion using intensive sampling data.• Pharmacokinetic profiles of MPA and MPAG with the UGT1A9 polymorphism in healthy Chinese were characterized. AIMSTo establish a population pharmacokinetic model that describes enterohepatic circulation (EHC) of mycophenolic acid (MPA) based on physiological considerations and to investigate the influence of polymorphisms of UGT1A9 on the pharmacokinetics of MPA. METHODSPharmacokinetic data were obtained from two comparative bioavailability studies of oral mycophenolic mofetil formulations. Nonlinear mixed effects modelling was employed to develop an EHC model including both MPA and its main glucuronide metabolite (MPAG) simultaneously. Demographic characteristics and UGT1A9 polymorphisms were screened as covariates. RESULTSIn total, 590 MPA and 589 MPAG concentration-time points from 42 healthy male volunteers were employed in this study. The chain compartment model included an intestinal compartment, a gallbladder compartment, a central and a peripheral compartment for MPA and a central compartment for MPAG. The typical population clearance (CL/F) estimates with its relative standard error for MPA and MPAG were 10.2 l h -1 (5.7%) and 1.38 l h -1 (6.9%), respectively. The amount of MPA recycled in the body was estimated to be 29.1% of the total amount absorbed. Covariate analysis showed that body weight was positively correlated with CL/F of MPA, intercompartment CL/F of MPA and distribution volume of MPA peripheral compartment. Polymorphisms of UGT1A9 did not show any effect on the pharmacokinetics of MPA and MPAG. The model evaluation tests indicated that the proposed model can describe the pharmacokinetic profiles of MPA and MPAG in healthy Chinese subjects. CONCLUSIONSThe proposed model may provide a valuable approach for planning future pharmacokinetic-pharmacodynamic studies and for designing proper dosage regimens of MPA.

Phytophthora root rot (PRR) of soybean (Glycine max (L.) Merr.) is the second most important cause of yield loss by disease in North America, surpassed only by soybean cyst nematode (Wrather et al. in Can J Plant Pathol 23:115-121, 2001). Tolerance can provide economically useful disease control, conditioning partial resistance of soybean to PRR. The aims of this study were to identify new quantitative trait loci (QTL) underlying tolerance to PRR, and to evaluate the effects of pyramided or stacked loci on the level of tolerance. A North American cultivar 'Conrad' (tolerant to PRR) was crossed with a northeastern China cultivar 'Hefeng 25' (tolerant to PRR). Through single-seed descent, 140 F2:5 and F2:6 recombinant inbred lines were advanced. A total of 164 simple sequence repeat (SSR) markers were used to construct a genetic linkage map. The percentage of seedling death was measured over 2 years (2007 and 2008) in the field at four naturally infested locations in Canada and China following additional soil infestation and in the greenhouse following inoculation with Phytophthora sojae isolate. A total of eight QTL underlying tolerance to PRR were identified, located in five linkage groups (F, D1b+w, A2, B1, and C2). The phenotypic variation contributed by the loci ranged from 4.24 to 27.98%. QPRR-1 (anchored in the interval of SSR markers Satt325 and Satt343 of LG F), QPRR-2 (anchored in the interval of Satt005 and Satt600 of LG D1b+w), and QPRR-3 (anchored in the interval of Satt579 and Sat_089 of LG D1b+w) derived their beneficial allele from 'Conrad'. They were located at chromosomal locations known to underlie PRR tolerance in diverse germplasm. Five QTL that derived beneficial alleles from 'Hefeng 25' were identified. The QTL (QPRR-1 to QPRR-7) that were detected across at least three environments were selected for loci stacking and to analyze the relationship between number of tolerance loci and disease loss percentage. The accumulation of tolerance loci was positively correlated with decreases in disease loss percentage. The pyramid of loci underlying tolerance to PRR provided germplasm useful for crop improvement by marker-assisted selection and may provide durable cultivar tolerance against the PRR disease.

The DDE model successfully described the non-linear pharmacokinetics of VPA. Furthermore, the proposed population pharmacokinetic model of VPA can be used to design rational dosage regimens to achieve desirable serum concentrations.