Several high-value decisions need to be taken when taking a compound from ‘bench’ to ‘bedside’. These could be related to the mechanism of action (e.g. is the target pathway inhibited sufficiently by the compound?) clinical management (e.g. What is the ideal trial design that optimizes dosing across multiple patient populations?) and interpretation of data from experiments or clinical trials (e.g. Does the slower than expected decay of the compound indicate additional physiological effects?).

We have worked on modeling and simulation projects in several therapeutic areas as well with research teams in early discovery all the way to design of clinical trials. We find that at the most basic level, modeling increases understanding of the connection between basic physiology and clinical outcomes. Large-scale models serve a knowledge-management function – a team’s knowledge (& hypotheses!) of physiology, mechanisms of action, clinical behavior etc. are all aggregated in one tool.”What if?” questions can be simulated readily to evaluate competing hypotheses and discrepancies in the data. We use simulations to evaluate between alternate experimental designs that can cost millions of dollars and recommend optimal course-of-action.

We use Quantitative Systems Pharmacology i.e. mechanistic models of physiology to gain insight into these questions. We extend these models to also study the effect of drugs on Virtual Patient populations and suggest optimum trial designs. We use simulations to evaluate between alternate experimental designs that can cost millions of dollars and recommend optimal course-of-action. Finally, we analyze the results of clinical trials and interpret the results in the light of the information gained from our models. This can be used to refine and improve models for further development.

Our current and past projects have been in several therapeutic areas such as sepsis, diabetes, dermatology, anemia, rheumatoid arthritis, hypertension. These projects can be associated with early discovery teams or clinical trial design in phase 3 or 4. We also develop tools for use in parameter estimation and optimization. This is an area which is very relevant to modeling of physiological systems since several parameters in quantitative models are not directly available and are estimated using indirect evidence.

Several high-value decisions need to be taken when taking a compound from ‘bench’ to ‘bedside’. These could be related to the mechanism of action (e.g. is the target pathway inhibited sufficiently by the compound?) clinical management (e.g. What is the ideal trial design that optimizes dosing across multiple patient populations?) and interpretation of data from experiments or clinical trials (e.g. Does the slower than expected decay of the compound indicate additional physiological effects?).

We have worked on modeling and simulation projects in several therapeutic areas as well with research teams in early discovery all the way to design of clinical trials. We find that at the most basic level, modeling increases understanding of the connection between basic physiology and clinical outcomes. Large-scale models serve a knowledge-management function – a team’s knowledge (& hypotheses!) of physiology, mechanisms of action, clinical behavior etc. are all aggregated in one tool.”What if?” questions can be simulated readily to evaluate competing hypotheses and discrepancies in the data. We use simulations to evaluate between alternate experimental designs that can cost millions of dollars and recommend optimal course-of-action.

We use Quantitative Systems Pharmacology i.e. mechanistic models of physiology to gain insight into these questions. We extend these models to also study the effect of drugs on Virtual Patient populations and suggest optimum trial designs. We use simulations to evaluate between alternate experimental designs that can cost millions of dollars and recommend optimal course-of-action. Finally, we analyze the results of clinical trials and interpret the results in the light of the information gained from our models. This can be used to refine and improve models for further development.

Our current and past projects have been in several therapeutic areas such as sepsis, diabetes, dermatology, anemia, rheumatoid arthritis, hypertension. These projects can be associated with early discovery teams or clinical trial design in phase 3 or 4. We also develop tools for use in parameter estimation and optimization. This is an area which is very relevant to modeling of physiological systems since several parameters in quantitative models are not directly available and are estimated using indirect evidence.

At Vantage, we are familiar with multiple tools to develop and analyse models and have used them in projects.

**PhysioLab**, **JDesigner** are tools to create Ordinary Differential Equation models using graphic interfaces that are accessible to both biologists and engineers. These tools facilitate development of models ‘from scratch’, modify existing models for simplification or expansion and to leverage existing assets in these platforms.

**R** is an open source statistical package used for statistics and graphical purposes. At Vantage, we use R for modeling, parameter estimation using Bayesian inference and for creating graphics of clinical data. R is available across all platforms and computationally intensive simulations can be performed by linking it to C, C++ and python codes.

**MATLAB** is proprietary computing software used in multiple applications. We have used Matlab to develop models as well as analysis, parameter estimation etc.

Systems approach to immuno-oncology (IO) drug development: Integrating Data and Knowledge

Modelling in Immuno-Oncology (IO) drug development: Perspective and challenges

QUANTITATIVE SYSTEMS PHARMACOLOGY: APPLICATIONS IN PHARMACEUTICAL R & D

QUANTITATIVE SYSTEMS PHARMACOLOGY CONCEPTS AND CHALLENGES

An open-source platform for Sensitivity Analysis of QSP models

Designing optimal basal insulin analogs using a Quantitative Systems Pharmacology model

Using optimization algorithms to estimate parameters in a simple PK model with bifurcations