AI-assisted Human Intestinal Absorption Prediction

With the increasing complexity and risk, human intestinal absorption (HIA) prediction is becoming more important in the drug discovery process. The drug needs to be dissolved, passed through the intestine, and finally spread or transported into the bloodstream. There is a growing need to understand and measure the effects of physicochemical properties of drugs on intestinal absorption processes. However, traditional experimental assays including in vitro and in vivo methods are expensive and time-consuming, which make it difficult to collect sufficient data to analyze structural contributions to the rate of intestinal absorption. To date, some in silico prediction methods, such as the quantitative structure-activity relationship (QSAR) model and machine learning techniques, have been developed to estimate the HIA of novel drugs with acceptable accuracy. Due to their extremely high throughput and low cost, in silico technologies have been seamlessly integrated into the drug discovery and development process. They provide valuable advantages, especially in ADMET analysis. Based on the advanced AI-assisted platform, Creative Biolabs can accurately predict the ADMET properties and help customers accelerate the drug screening process and reduce R&D costs.

Fig.1 Models for predicting human intestinal absorption. (Price, et al., 2021)

Human Intestinal Absorption (HIA)

The high rate of attrition in drug discovery is mainly due to the absorption, distribution, metabolism, and excretion characteristics of the candidate compounds. Many active drugs fail during phase II or III clinical development because of their poor ADME performance. Therefore, the study of absorption, distribution, metabolism, and elimination must be carefully considered in the process of drug discovery, and experimental data obtained through high-throughput screening to pursue better ADME properties. HIA is one of the most important ADMET properties and a key step during the drugs’ transporting to their targets. The utilization of drugs in the human body is a very complex process that is difficult to accurately analyze using statistical models. Furthermore, it is also difficult to predict the oral bioavailability of various drugs because of the variety of components that play a role in this process. A useful prediction model for human oral bioavailability must have strong characteristics linked to carrier-mediated transport and first-pass metabolism due to the variety of drug absorption pathways. And as HIA is one of the key factors that affect bioavailability, great effort has been put into developing an effective HIA prediction method.

Computational Prediction of HIA

Pharmacokinetic information of compounds is an important part of drug design and development. Identification of the factors affecting the absorption, distribution, metabolism, and excretion of the compounds is necessary for modeling the pharmacokinetic properties. The uptake of drug compounds by human intestinal cells is an important property of potential drug candidates. However, the measurement of this property is costly and time-consuming. In silico methods like the QSAR model is an attractive alternative to experimental measurements to estimate the percent human intestinal absorption (%HIA). Extrapolation from characteristic compounds to untested molecules seems feasible under the assumption that similar structures exhibit similar biological activities. Various statistical methods can be applied to create more or less predictive models from molecular data, including linear regression and machine learning (e.g., support vector machines (SVM) or artificial neural networks (ANN)). Currently, a large number of individual predictive models are available for absorption study using machine learning approaches. Additionally, in order to reduce the rate of attrition of drug candidates entering preclinical and clinical trials, there are ongoing attempts to use various artificial intelligence methods to predict the intestinal absorption of compounds.

Fig.2 Results of prediction algorithms. (Kumar, et al., 2017)

As a reliable partner to the world's leading pharmaceutical companies and research institutions, Creative Biolabs brings together the market-leading in silico expertise to build our AI-assisted platform that can provide the professional human intestinal absorption prediction.

References

1. Price, E.; et al. Global analysis of models for predicting human absorption: QSAR, in vitro, and preclinical models. Journal of Medicinal Chemistry. 2021, 64(13): 9389-9403.
2. Kumar, R.; et al. Prediction of human intestinal absorption of compounds using artificial intelligence techniques. Current drug discovery technologies. 2017, 14(4): 244-254.