Skip links

Our Knowledge-Empowered Modeling Platform

Differentia Bio’s platform integrates AI & Data Analytics, Knowledge, and Modeling. Our AI & ML analytics, trained by our proprietary data pool, empowers our modeling & simulation services, and enables us to deliver knowledge-based insights and optimize drug development for a greater success rate.

Our Holistic Platform Unites Knowledge, AI & Data Analytics, and Modeling

Knowledge

Our platform’s foundation rests on a comprehensive understanding of biological systems derived from extensive research, collaborations with experts, leading universities, and hospitals. This knowledge repository encompasses detailed information about biological pathways, molecular interactions, disease mechanisms, pharmacological and pathology insights. It’s continually updated to ensure accuracy and scientific excellence. Embedded in our modeling platform, it enables us to prioritize targets, identify biomarkers and test combination therapies hypothesis.

AI & Data Analytics

Empowered by our proprietary multi-omics patient datasets and our in-house knowledge graphs that integrate drug, disease, protein and immune system databases, we leverage cutting-edge Artificial Intelligence (AI) and Machine Learning (ML) for advanced data analytics. Our ML empowered models support the development of tailored and detailed mechanistic models, enable parameter inference, and enhance the decision-making process. This approach enables the identification of hidden insights, patterns and correlations, and the rapid development of detailed QSP models.

Modeling

Our Knowledge-Empowered Modeling platform supports all stages of drug development, enabling:

Discovery
  • Target identification and validation
  • Drug profile identification and benchmarking
  • Modeling support for a variety of therapeutic modalities spanning the range from small molecules to biologics and cell and gene therapies
Pre-Clinical
  • Candidate selection
  • Therapeutic index assessment
  • Preclinical to clinical translation and clinical dose selection
Clinical
  • Exposure-response in safety and efficacy
  • Biomarker discovery
  • Dose predictions for Ethnicity bridging, Paediatric extrapolation & Special populations with renal or hepatic impairment, Virtual patient populations and Digital Twins

Services We Provide

DISCOVERY

Target Selection and Validation

By allowing researchers to hypothesize and test new ideas in silico, detailed disease models can accelerate candidate target identification by focusing research efforts on the most clinically relevant pathways involved in disease progression. Also, by defining the appropriate laboratory experiments to empirically test those hypotheses, the process of target validation becomes more efficient. These models can identify pathways and targets having the greatest clinical impact, systematically prioritize dozens of targets and pathways for comparative analysis, formulate and test hypotheses regarding the mechanism of action for prospective therapeutic targets and design appropriate confirmatory experiments and evaluate the potential function of novel genes and determine the in vitro/in vivo experiments needed to confirm those predicted behaviors

Best in class properties and Competitor benchmarking

Models can address species-specific changes in pathways of interest as well changes that may occur in pathways due to adaptation resulting from a chronic disease. This allows a biopharmaceutical company to focus on lead optimization using the meaningful representations of the pathway of interest against the most clinically relevant behavior. Candidate molecules can be analyzed against known therapies to compare them under a wide range of clinical protocol environments, allowing better lead optimization, better clinical development optimization, or both.

Researchers can calculate, based on the desired clinical effects, what the ideal in vitro (tissue, cell) and in vivo (animal) characteristics would be for a drug, including its pharmacological and pharmacokinetic properties. Armed with such information, companies can select the most appropriate studies and experimental conditions to focus their data collection efforts and set explicit optimization (activity/pharmacokinetic) objectives for lead optimization cycles. Such models can allow one to evaluate how differences in preclinical measurements (e.g., PK absorption and clearance rates, Ki values, receptor specificity, etc.) affect clinical outcome for a given candidate or a series of lead candidates, evaluate multiple pathways where the lead candidate acts, supporting lead optimization in chemistry by maximizing sensitive mechanisms and minimizing non-sensitive mechanisms effecting clinical endpoints, given a specific target, define what therapeutic thresholds (Ki values, plasma concentration, etc.) are necessary for a lead candidate to exhibit clinical efficacyand compare against clinical competitors in the same simulation environment and evaluate best-in-class properties

PRECLINICAL TO CLINICAL TRANSLATION

Biomarker Identification

Biomarkers can serve as clinical endpoints to patient response or as surrogate endpoints in drug discovery. Modeling can be used to identify cells and pathways that have a critical impact on clinical outcome and identify temporal changes that are consistent and predictive of clinical efficacy. Thus, we can identify biomarkers and provide a biological context for understanding their mechanism of action.

First-in-human dosing and therapeutic index prediction

Differentia builds models that integrate all the in vitro and in vivo data generated in a project to predict FIH dose, efficacious dose and the therapeutic index. Models allow researchers to translate in vitro/in vivo efficacy data into a MABEL dose as well as translate safety data from animal studies into NOAEL predictions in humans under different routes of administration and dosing regimens and support the IND submission.

CLINICAL TRIAL PLANNING AND OPTIMIZATION

Dose escalation and expansion

As data is generated in clinical trials the patient response can be captured in PK/ER models. In addition, models can be individualized to capture the variability seen in the pharmacokinetics and pharmacodynamics of the response. Population PK and population PD models, virtual patient cohorts and PK/ER models allow for the planning of trials in Phase II and Phase III. These models assist in understanding the relationship between the expected response and patient covariates and the overall variability that can be expected in a population. These insights can be used to power trials appropriately and identify the right data to collect along with the best collection times.

Biomarker discovery and validation

Surrogate markers or biomarkers identified in the discovery process can be validated by comparing their levels in clinical samples to expected changes predicted by modeling. In addition, de novo markers can be identified from the clinical samples by developing correlations between their dynamic changes and clinical responses either statistically or using AI/ML methods

Explore
Drag