The basis of our technology is a mechanistic, systems biologic approach: Thousands of biochemical reactions are represented mathematically and implemented into computer-based network models, the Insilico Cells™. The spectrum of considered pathways comprises metabolic reactions as well as gene regulation and signal transduction reactions.

Following customisation and verification of the Insilico Cells™ through implementation of our customers' data, Insilico is able simulate numerous cellular scenarios systematically by means of high performance computing and proprietary software. This allows for the quantification and prediction of the effects of variations of the genetic background of the cells as well as of changes in external conditions affecting the cells, e.g. nutrient or drug concentrations. Thus expectable beneficial effects can be weighed against potential expectable side effects so that Insilico can support its customers by making reasoned recommendations.

In addition to its software applications Insilico Inspector™ and Insilico Discovery™, which can be licensed by customers, Insilico utilises further software platforms, databases and high-end computer architectures to simulate biological reactions.

Insilico Simulation Platform Rely on a rich toolbox for a wide range of simulation applications

Apart from the simulation modules implemented in Insilico Discovery™ comprising e.g. Metabolic Flux Analyses and dynamic simulations of differential-algebraic systems, the Insilico Simulation Platform features further simulation tools (excerpt):

The Gene Perturbation Screening Tool allows users to screen millions of combinations of up to ten different genes for perturbations leading to enhanced product titre and productivity. This tool not only considers the impact of gene perturbations on product formation but also its impact on cell growth and by-product formation, thus allowing goal-oriented rational strain development.

The Tool for the Adaptation of Feed Compositions and Feed Strategies facilitates the identification of over or underfeeding scenarios prior to the start of fermentation, thus saving valuable process development time.

With the Pathway Screening Tool, Insilico provides the possibility to simulate new or alternative pathways in a host cell in parallel. For this purpose, the tool considers more than 8,000 biochemical reactions contained in Insilico's reaction database. Moreover, the tool is able to integrate new reactions bearing resemblance to known reactions. Hence the effort needed to adapt a host cell to produce a new product can be predicted. Thus the potential and risks of a new project can be assessed, and novel Intellectual Property can be created.

By direct coupling of Insilico's Simulation Platform to Insilico's Optimisation Platform, a high degree of automation and hence speed-up of the workflow is achieved.

Insilico Optimisation Platform Find the optimal development strategy fast and reliably

Insilico’s optimisation platform is capable of solving (i) convex as well as (ii) non-linear, non-convex optimisation problems according to high industrial standards.

In the area of convex optimisation, we apply e.g. linear optimisation for Flux Balance Analysis (FBA) and quadratic optimisation for Minimisation of Metabolic Adjustment (MOMA).

In the area of non-linear, non-convex optimisation, we utilise a parallel openMPI version of pCMAlib, a FORTRAN library for the Covariance Matrix Adaptation Evolution Strategy (CMA-ES).

This high-performance optimisation platform allows Insilico's experts to utilise cluster resources efficiently and to solve optimisation problems with up to several thousands of parameters. Its scalability enables users to request additional resources easily if optimisation problems become larger and optima become harder to find.

Insilico uses its high-performance optimisation platform to identify the parameters of dynamic network models, to find the optimal feed strategy and composition, to identify intracellular flux distributions from 13C labelling experiments, and for a variety of other applications.

Insilico Databases Consider more than 8,000 biochemical reactions at one go

Insilico offers ready-to-use network models for more than 20 microbial, mammalian and insect cell strains. Over 2,000 biochemical compounds listed in a central compound database can be implemented into these models. These compounds, in turn, are involved in more than 5,000 enzymatic and more than 3,000 non-enzymatic reactions from more than 60 different organisms, which together form Insilico's reaction database. These databases are steadily expanded and updated, providing the basis for Insilico's portfolio of modelling and simulation services.

Insilico High Performance Computing Exploit Europe's fastest civil computer to address your questions

Cutting computing time dramatically is a building block for shortening the time to market in various development processes. Insilico conducts simulations and optimisations of utmost complexity which need computing power far beyond the capacity of a customary personal computer. For this reason, Insilico utilises high performance computer clusters with the aid of its proprietary software tools.

Accessing Europe’s very fastest civil computer named "Hermit", Insilico has entered the dimension of petaflop computing, which means dealing with more than one quadrillion computing operations per second. This is highly beneficial, especially for the design of innovative whole-body models. Thanks to its massively parallel architecture, the new computer is ideal for running the simulation of many different cells at the same time without being constrained by retarding load peaks.

The seamless interoperability of Insilico Discovery, Insilico's licensable simulation software, with Insilico's parallel optimisation platform allows our customers to utilise the computing power of high performance computers using an easy-to-use graphical user interface. The object-oriented representation of a cellular network model in Insilico Discovery is automatically translated into FORTRAN code. The generated FORTRAN code is then transferred to the supercomputer, compiled and started on the supercomputer. Depending on the complexity of the optimisation, an appropriate amount of computing resources is allocated for an appropriate amount of time. While optimisation is running, intermediate results can be reviewed at any time. The final result is transferred back to the user and translated into the object-oriented representation of the cellular network model. This completely automated optimisation workflow allows users to exploit supercomputer resources without having to deal with complex technical details.