Biology subjects

Cordes, H.

Publications and source records attributed to Cordes, H..

4 recordsLinked to original sources

bioRxiv2025

Towards pharmacokinetic profile predictions for monoclonal antibodies using sequence based machine learning derived parameters and compartmental modeling

This study presents a novel approach for predicting complete pharmacokinetic (PK) profiles of monoclonal antibodies (mAbs) based solely on their amino acid sequences, addressing a critical gap in early-stage therapeutic antibody development. While current methodologies rely on in vivo testing in specialized models like Tg32 mice or cynomolgus monkeys, our approach enables PK parameter prediction prior to molecule synthesis. Using a dataset of 118 diverse mAbs, we developed a physics-informed neural network (PINN) incorporating a two-compartment PK model. The four parameters of the PK model are predicted by a NN in which the numerical representation of the light and heavy chain sequences derived through sequence-attention and protein language model features serve as input. The model successfully predicted clearance and volume of distribution at steady state with remarkable accuracy, achieving predictions within 2-fold error for 10/13 and 13/13 mAbs in the test set, respectively. Concentration-time profiles were well-characterized with a median geometric mean fold error (GMFE) close to 2, with nearly 70% of predictions having GMFE less than 2. This proof-of-concept study demonstrates the feasibility of sequence-based PK prediction for mAbs, potentially reducing animal testing requirements while accelerating candidate selection in therapeutic antibody development. The approach aligns with regulatory initiatives promoting new approach methodologies and complements existing sequence-based predictions of protein structure and target binding during the design phase.

systems biology↗

bioRxiv2025

Machine Learning Meets Pharmacokinetics: A Comparative Analysis of Predictive Models for Plasma Concentration-Time Profiles

Predicting pharmacokinetic (PK) profiles from molecular structures constitutes a significant advancement in pharmaceutical research with substantial implications for expediting the entire drug discovery process. Our investigation presents a comprehensive comparative analysis of five distinct methodological frameworks for predicting rat plasma concentration-time profiles: four approaches that integrate mechanistic modeling with computational techniques, and one approach employing pure machine learning (ML) architecture without physiological constraints. As reference (1), we used a method that predicts non-compartmental analysis (NCA) parameters to predict plasma PK profiles with a one-compartmental PK model (NCA-ML). Alternative approaches include: (2) a physiologically based PK (PBPK) model utilizing ML-predicted in vitro characteristics; (3) CMT-ML, where neural networks predict compartmental PK model parameters; (4) CMT-PINN, employing physics-informed neural networks; and (5) PURE-ML, using decision trees to predict concentration values at specific timepoints based on predicted volume of distribution and clearance. Assessment using multiple performance metrics demonstrated that the CMT-PINN approach achieved superior predictivity for PK profiles. Models trained directly on concentration-time, rather than derived PK parameters, delivered markedly improved predictive performance, particularly when working with limited training datasets. Our findings confirm the viability of predicting PK behavior from molecular structures before synthesis, representing a significant advancement for subsequent design. Implementation of these computational approaches enables informed compound selection early in the discovery pipeline, concentrating valuable resources for advanced preclinical investigations on only the most promising candidates, with anticipated beneficial effects on overall development timelines.

systems biology↗

bioRxiv2025

Talk2Biomodels: AI agent-based open-source LLM initiative for kinetic biological models

In this study, we present Talk2Biomodels (T2B), an open-source1, user-friendly, large language model-based agentic AI platform designed to democratize access to computational models of biological interactions and promote the FAIRification (Findability, Accessibility, Interoperability, and Reusability) of these models. T2B enables users to explore and analyse mathematical models of biological systems using natural language. It eschews the traditional graphical user interface (GUI) and minimally adaptable workflow in favour of a modern agentic framework to provide a dynamic and immersive experience to explore mathematical models of biological interactions through conversations in natural language. T2B supports models encoded in the open-source community format Systems Biology Markup Language (SBML) and is integrated with the BioModels database (https://www.ebi.ac.uk/biomodels/), enabling seamless exploration, simulation, and analysis of curated systems biology models. Use cases in precision medicine, epidemiology, and emergent systems properties of biological networks are presented to demonstrate how experts and non-experts in computational biology can benefit from T2B.

systems biology↗

bioRxiv2018

A physiology-based model of bile acid metabolism shows altered tissue concentrations after drug administration and in specific genotype subgroups

Drug-induced liver injuries (DILI) are an important issue in drug development and patient safety and often lead to termination of drug-development programs or late withdrawals of drugs. Since DILI events are hard to diagnose in preclinical settings, a need for alternative prediction methods such as computational modeling emerges. Impairment of bile acid (BA) metabolism, known as cholestasis, is a frequent form of DILI. Being rather a systemic then a single organ related disease, whole-body physiology-based modeling is a predestined approach for cholestasis modeling. The objectives of the presented study were 1) the development of a physiology-based model for human bile acid metabolism, 2) model validation and characterization for a virtual population, and 3) prediction and quantification of the effects of genetic predispositions and drug interaction on bile acid metabolism. The developed physiology-based bile acid (PBBA) model is based on the standard PBPK model of PKSim(R) and describes the bile acid circulation in a healthy reference individual. Active processes such as the hepatic synthesis, gallbladder emptying upon meal intake, transition through the gastrointestinal tract, reabsorption into the liver, distribution within the body, and excretion are included. The kinetics of active processes for the surrogate BA glycochenodeoxycholic acid were fitted to time-concentration profiles of blood BA levels reported in literature. The robustness of our PBBA model is underlined by the comparison of simulated plasma BA concentrations in a virtual population of 1,000 healthy individuals with reported data. In addition to plasma concentrations, the PBBA model allows simulations of BA exposure in relevant tissues like the liver and can therefore enhance the mechanistic understanding of cholestasis. This feature was used to analyse the reported increased risk of cholestatic DILI in Benign Recurrent Intrahepatic Cholestasis type 2 (BRIC2) patients. Simulations of the PBBA model suggest a higher susceptibility of BRIC2 patients towards cholestatic DILI due to BA accumulation in hepatocytes. Apart from these intrinsic effects, drug-interactions and their effect on the systemic bile acid metabolism were simulated by combining the PBBA model with a drug PBPK model of cyclosporine A (CsA). The results of which confirmed the reported higher risk of developing DILI as a consequence of CsA intake. Altogether, the presented model enhances our mechanistic understanding of cholestasis, allows the identification of drug-interactions leading to altered BA levels in blood and organs, and could be used to prevent clinical cases of cholestasis and enhance patient safety.

systems biology↗