We are building frontier virtual modeling and simulation technologies that are loved by biologists of all kinds, from classical to computational.
Decoding biology and understanding how it works is one of the future frontiers in scientific research. The potential is enormous, from curing disease to climate resistant crops and manufacturing superior biological materials.
Yet, we are still far away from illuminating the inner workings of biological cells and their interactions. This is because biology is a dynamic, non-linear system of proteins, enzymes, metabolites, molecules and cells interacting with each other. This system is constantly evolving and responding to perturbations, changes and external factors.
For example, an average of 86% of drugs fail in clinical trials and it costs USD 2 billion to bring a drug to the market. The key reasons for this high failure rate is a lack of understanding of the underlying (disease) biology and the mechanism of action of the drug on a patient level.
It currently takes months to years and a team of mathematicians or physicists and biologists and medical professionals to build large mechanism-based models because accessible tools and framework are still missing. This is why we design intuitive, web-based user interfaces for scientists to build and collaborate on models. Database integrations and advanced AI tools will make this a seamless, semi-automated experience.
To simulate these models, we are building the world’s most powerful and versatile simulation engine, the EndoGenics™ Simulation Engine. We developed from first principles a framework to integrate different biological mechanisms and laws and evolve them over time. This framework is able to handle highly non-linear dynamics and extreme variation in rates, often observed in biology between different mechanisms. As such, it is the first simulation engine natively able to simulate models built from common data structures, like the pathway database Reactome. It is further designed to be suitable for modern high performance computing architectures to deal with large scale models and simulations.
A core component of this infrastructure is its interface to computational biology. Artificial intelligence and machine learning are tremendous tools to extract patterns and trends from large data sets, mostly omics data, which are currently proliferating at massive scale throughout the life sciences with terabytes of data per experiment. Signals and statistical insights identified from big data will direct and inform model building and are the basis for model calibration and optimization.
Computational and classical biologists need to work hand in hand to progress our understanding of biology. This is exactly what we enable.
Currently, accessible tools and frameworks are missing to build large scale mechanism based models. To do so, three fundamental challenges need to be addressed:
We have solved points 1 and 2 with a framework that is ready to integrate key mechanisms including:
With the EndoGenics™ Simulation Engine and through democratizing access to it with best in class user interfaces and experiences to build, simulate, visualize and analyse models and their simulations we accelerate the journey towards solving the third challenge: filling the knowledge gaps.
Efforts are currently underway to build the AI Virtual Cell. Despite the tremendous capability of machine learning to identify patterns, signals and trends, it lacks causal explainability and is constrained by the data it is trained on. In particular, when applied to temporal evolution, like disease trajectories, these models cannot predict future states, except statistically.
A complementary, hybrid approach is necessary, leveraging the capabilities of machine learning to identify patterns and trends from large amounts of data and combining it with mechanism based models for causal explainability to generate new biological knowledge.
With implementing the framework and infrastructure for large scale mechanism based and AI enhanced models, building the virtual whole cell comes a large step closer to reality.