Computational systems biology is a research field that combines mathematical and computational models together with molecular data in order to gain a better understanding of various biological systems. Among many different problems, it has also been applied to the study of cancer in an attempt to understand the behavior of tumors and predict the efficacy of treatment schemes. However, modelling and simulating cancer tumors is a challenging problem because of the multi-scale nature of this complex multicellular disease. A typical in-silico experiment may result in tens of thousands of simulations being produced by the underlying model that need to be analyzed, even through simple visual inspection, in order to determine which of them might actually be promising from a therapeutic point of view. As a result, statistical and machine learning techniques are employed to analyze and categorize the ouput of such in silicon experiments. As a next step, such methods may also be used to predict whether a simulation modelling a tumor and its treatment via a drug will turn out to be efficacious or not, thus saving valuable computational resources. The goal of this project is to investigate the effectiveness of statistical (e.g., time-series analysis) and machine learning (e.g., deep learning approaches, such as LSTM networks) methods (and even additional forecasting methods) on the task of predicting the outcome of such tumor simulations running on High-performance computing clusters.
Indicative bibliography:
• Evgenios Kladis, Charilaos Akasiadis, Evangelos Michelioudakis, Elias Alevizos, Alexandros Artikis, An Empirical Evaluation of Early Time-Series Classification Algorithms. EDBT/ICDT Workshops 2021 http://cer.iit.demokritos.gr/publications/papers/2021/ETS_Paper.pdf
• Charilaos Akasiadis, Miguel Ponce de Leon, Arnau Montagud, Evangelos Michelioudakis, Alexia Atsidakou, Elias Alevizos, Alexander Artikis, Alfonso Valencia, Georgios Paliouras, Parallel Model Exploration for Tumor Treatment Simulations. CoRR abs/2103.14132 (2021) https://arxiv.org/pdf/2103.14132
Contact: cakasiadis [at] iit.demokritos.gr, alevizos.elias [at] demokritos.gr