Predicting Fraction Unbound in Human Plasma from Chemical Structure: Improved Accuracy in the Low Value Ranges

Abstract

Predicting the fraction unbound in plasma provides a good understanding of the pharmacokinetic properties of a drug to assist candidate selection in the early stages of drug discovery. It is also an effective tool to mitigate the risk of late-stage attrition and to optimize further screening. In this study, we built in silico prediction models of fraction unbound in human plasma with freely available software, aiming specifically to improve the accuracy in the low value ranges. We employed several machine learning techniques and built prediction models trained on the largest ever data set of 2738 experimental values. The classification model showed a high true positive rate of 0.826 for the low fraction unbound class on the test set. The strongly biased distribution of the fraction unbound in plasma was mitigated by a logarithmic transformation in the regression model, leading to improved accuracy at lower values. Overall, our models showed better performance than those of previously published methods, including commercial software. Our prediction tool can be used on its own or integrated into other pharmacokinetic modeling systems.

Development of an in silico prediction system of human renal excretion and clearance from chemical structure information incorporating fraction unbound in plasma as a descriptor

“…S4, a correlation could be seen between observed and predicted f u,p values ( r = 0.84), with 72.8% and 84.0% of the predicted f u,p values falling within 2-fold and 3-fold error, respectively. This indicated that the f u,p predicted by f u,p predictor 22 correlated well with the observed f u,p .…”

“…The compounds that displayed CR < 0.67, 0.67 ≤ CR < 1.5, or 1.5 ≤ CR were classified into reabsorption (R) type (net reabsorbed compounds), intermediate (IM) type (apparently not reabsorbed or secreted compounds), and secretion (S) type (net secreted compounds), respectively 5 . Predicted f u,p was calculated using our previously developed f u,p predictor 22 . Ionization profiles in the data set were extracted from the ChEMBL database.…”

“…Previously, we constructed prediction models of the human unbound fraction in plasma ( f u,p ) 22 , with the f u,p prediction models released via a freely available tool ( f u,p Predictor, http://adme.nibiohn.go.jp/fup/). As approximately 10% of the blood volume is filtered at the glomerulus by the hydraulic pressure exerted by the arterial blood and, as a general rule, only the unbound drug in plasma is filtered, the value of f u,p significantly impacts the renal glomerular filtration 23 .…”

“…Moreover, f u,p has been included as one of the four default descriptors in f e prediction in several reports 12–14 . Thus, we considered that our f u,p prediction models 22 might be expanded to predict f e and CL r .…”

Predicting the Acute Liver Toxicity of Aflatoxin B1 in Rats and Humans by an In Vitro–In Silico Testing Strategy

“…This value was derived in silico using the fu,p Predictor software. [22] For rats, the f ub, blood was considered the same as for humans. Since the PBK models predict the AFB1 concentration in blood, for evaluation of the model performance, the plasma concentrations of AFB1 from in vivo kinetic studies were converted to blood concentrations to enable comparison to the model predictions, assuming that blood concentrations are 0.6 and 0.55 times the plasma concentrations in rats and humans, respectively.…”