Periodic Reporting for period 5 - CanPathPro (Generation of the CanPath prototype - a platform for predictive cancer pathway modeling)
Período documentado: 2020-03-01 hasta 2021-11-30
Cancer is the result of mutations and other alterations in multiple genes, which give rise to the disease’s enormous molecular complexity affecting multiple signalling pathways and their cross talk. An influx of tumour “omics” data (genomic, transcriptomic, proteomic) has allowed the profiling of this disease in unprecedented detail. Nonetheless, the challenge remains to translate this knowledge into clear benefits for better treatment, drug development and for an enhanced understanding of the molecular basis and progression of cancer.
Genetically engineered cancer mouse models are one of the main tools for functional analysis of cancer alterations. The influx of tumour omics data has increased the need for new mouse models that recapitulate the newly discovered molecular profiles of the human disease. However, pragmatic limitations in financial and time resources preclude massively parallel experimentation, thus slowing down the progress of discovery. In the meantime, omics datasets are increasing, descriptive of numerous health and disease states, but nonetheless remain underexploited.
The main bottleneck in these efforts is a lack of efficient, validated tools which could integrate and analyse the omics datasets. Current approaches are mainly confined to statistical analysis, molecular pattern recognition through machine learning or -at best- modelling of single pathways. These approaches do not consider the complex pathways and their cross-talk, which ultimately determines cancer initiation, progression and drug response.
CanPathPro has addressed these challenges. The overall objective of the project was to build and validate a combined experimental and systems biology platform, to be utilised in testing and generating new cancer signalling hypotheses, in biomedical research. It has combined, in a single platform, omics and quantitative immunohistopathological data of cancer mouse models with analytical, modelling, predictive and visualisation computational tools. The platform performs data integration and predictive modelling (i.e. in silico predictions based on computational and mathematical modelling using large-scale datasets) of the relevant signalling networks, leading to an output of testable hypotheses.
The components that have been used in the development of the CanPathPro platform comprise highly defined mouse and organotypic experimental systems, next generation sequencing, SWATH-based proteomics and a systems biology computational model for data integration, visualisation and predictive modelling. The project has taken an unique approach, combining classic cancer research with omics data and computational modelling to develop and validate a new biotechnological application: a platform for generating and testing cancer signalling hypotheses in biomedical research.
Genetically engineered cancer mouse models are one of the main tools for functional analysis of cancer alterations. The influx of tumour omics data has increased the need for new mouse models that recapitulate the newly discovered molecular profiles of the human disease. However, pragmatic limitations in financial and time resources preclude massively parallel experimentation, thus slowing down the progress of discovery. In the meantime, omics datasets are increasing, descriptive of numerous health and disease states, but nonetheless remain underexploited.
The main bottleneck in these efforts is a lack of efficient, validated tools which could integrate and analyse the omics datasets. Current approaches are mainly confined to statistical analysis, molecular pattern recognition through machine learning or -at best- modelling of single pathways. These approaches do not consider the complex pathways and their cross-talk, which ultimately determines cancer initiation, progression and drug response.
CanPathPro has addressed these challenges. The overall objective of the project was to build and validate a combined experimental and systems biology platform, to be utilised in testing and generating new cancer signalling hypotheses, in biomedical research. It has combined, in a single platform, omics and quantitative immunohistopathological data of cancer mouse models with analytical, modelling, predictive and visualisation computational tools. The platform performs data integration and predictive modelling (i.e. in silico predictions based on computational and mathematical modelling using large-scale datasets) of the relevant signalling networks, leading to an output of testable hypotheses.
The components that have been used in the development of the CanPathPro platform comprise highly defined mouse and organotypic experimental systems, next generation sequencing, SWATH-based proteomics and a systems biology computational model for data integration, visualisation and predictive modelling. The project has taken an unique approach, combining classic cancer research with omics data and computational modelling to develop and validate a new biotechnological application: a platform for generating and testing cancer signalling hypotheses in biomedical research.
The work performed in this project has provided the foundation for the generation of the CanPathPro prototype. Mouse cancer models have been generated. Tumour samples were collected and analysed on both genetic (DNA & RNA) and proteomic (quantitative proteome and phospho-proteome) levels. The computational model has been further developed, improved and validated experimentally. Computational methods for model simulation, parameter estimation, uncertainty and identifiability analysis have been improved and extended while approaches for visualization of data and model systems have been developed as basis for the CanPathPro prototype. The system has been tested in the context of academic use for the generation of new hypothesis as proof-of-principle for a commercial platform. Using these models, we were able to build up a clearer picture of the effect of specific molecular alterations on cell signalling networks, using predictive modelling to generate hypotheses that were validated experimentally. Future exploitation of the platform has been investigated with the development of business and commercialisation plans. CanPathPro’s visibility and communication efforts have been heightened, with an online and social media presence, communication materials (see our Explainer video – www.canpathpro.eu/news/) and a large body of peer reviewed publications.
CanPathPro has been developing and validating a bioinformatics concept and the associated computational tools required for the translation of highly complex and heterogeneous omics data into predictive modelling of cancer signalling.
The modelling system at the centre of CanPathPro – ModCellTM, a large mathematical model of cancer-related signal transduction pathways represents progress beyond the state of the art, because it provides exceptionally high-level signal transduction analysis (incorporating the functional effects of perturbations on components of thousands of biochemical reactions and identifying cross-talk between pathways) on a quantitative basis. In addition, the extensive refinement of the computational model (via parameter optimisation as well as iterative rounds of in silico modelling and in vivo validation) represents progress beyond state-of-the-art among systems biology approaches delineating cancer signalling.
The main result of the project is the CanPathPro prototype: A tripartite tool for the (i) generation and (ii) integration of quantitative mouse omics data, followed by (iii) predictive in silico modelling of cancer signalling pathways and networks in mouse models. The CanPathPro prototype has been built as a commercial platform; its development represented the main aim of the project. Additional results include the delineation of breast and lung signalling pathways, which have been analysed in CanPathPro during model development, and the optimisation of associated methodologies (organotypic systems, omics platforms, HCA assays, etc.) relevant for cancer research.
Robust and accurate functional predictions of the behaviour of complex biological systems, based on a validated system will have broad and significant impact on diverse areas: from cancer research and personalised medicine to drug discovery and development. The in silico modelling and high-performance computing tools will provide completely new solutions for researchers, SMEs and industry, in interpretation and analysis of omics data. Thus, in the medium to long run, the output of the project will contribute significantly to improving outcomes for cancer patients, beyond breast and lung cancer sufferers. Additional impacts include bridging of knowledge from different disciplines (functional genomics, cancer biology, oncology, mathematics, systems biology); generation of new hypotheses for diagnosis and treatment of cancer and their in silico testing, by rapid simulation of millions of experimental conditions. The latter can facilitate prioritisation of experimental validation, effectively reducing animal experiments while enhancing the efficient use of time and financial research resources.
The modelling system at the centre of CanPathPro – ModCellTM, a large mathematical model of cancer-related signal transduction pathways represents progress beyond the state of the art, because it provides exceptionally high-level signal transduction analysis (incorporating the functional effects of perturbations on components of thousands of biochemical reactions and identifying cross-talk between pathways) on a quantitative basis. In addition, the extensive refinement of the computational model (via parameter optimisation as well as iterative rounds of in silico modelling and in vivo validation) represents progress beyond state-of-the-art among systems biology approaches delineating cancer signalling.
The main result of the project is the CanPathPro prototype: A tripartite tool for the (i) generation and (ii) integration of quantitative mouse omics data, followed by (iii) predictive in silico modelling of cancer signalling pathways and networks in mouse models. The CanPathPro prototype has been built as a commercial platform; its development represented the main aim of the project. Additional results include the delineation of breast and lung signalling pathways, which have been analysed in CanPathPro during model development, and the optimisation of associated methodologies (organotypic systems, omics platforms, HCA assays, etc.) relevant for cancer research.
Robust and accurate functional predictions of the behaviour of complex biological systems, based on a validated system will have broad and significant impact on diverse areas: from cancer research and personalised medicine to drug discovery and development. The in silico modelling and high-performance computing tools will provide completely new solutions for researchers, SMEs and industry, in interpretation and analysis of omics data. Thus, in the medium to long run, the output of the project will contribute significantly to improving outcomes for cancer patients, beyond breast and lung cancer sufferers. Additional impacts include bridging of knowledge from different disciplines (functional genomics, cancer biology, oncology, mathematics, systems biology); generation of new hypotheses for diagnosis and treatment of cancer and their in silico testing, by rapid simulation of millions of experimental conditions. The latter can facilitate prioritisation of experimental validation, effectively reducing animal experiments while enhancing the efficient use of time and financial research resources.