You are here: TUCS > EDUCATION > TUCS Short Courses > Joint TUCS-BioCity One-Day Tutorial

## Joint TUCS-BioCity One-day Tutorial on Modern Computational Modelling and GPU-Based Simulation for Life Sciences (2 ECTS)

On Monday 21.11.2016 TUCS and BioCity organise a joint one-day tutorial on modern computational modelling and GPU-based simulation for life sciences (2 ECTS).

The event is organised through TUCS’s research program on BioHealth and BioCity’s research program on Computational and Molecular Methodologies for Life Sciences. The guest lecturers are Daniela Besozziand Marco Nobile, both from University of Milan-Bicocca, Italy.

The seminar takes place in room XXII, Agora.

The schedule of the seminar is as follows:

- 09.00 - 10.30: Limits and strengths of mathematical modeling for biological
systems (Daniela Besozzi)

[Download the slides as PDF here.] - 10.30 - 11.00: coffee break
- 11.00 - 12.30: Reaction-based modeling: hands-on session (Daniela Besozzi)

[Download the slides as PDF here.] - 12.30 - 13.30: lunch break
- 13.30 - 15.30: Graphics Processing Units for Life Sciences (Marco Nobile)

[Download the slides as PDF here.] - 15.30 - 16.00: coffee break
- 16.00 - 17.00: GPUs for Systems Biology, in practice (Marco Nobile)

Abstracts of the talks and short bios of the guest lecturers are given below.

### Limits and strengths of mathematical modeling for biological systems & reaction-based modeling

Mathematical modeling of biological systems generally pose to modelers questions like: “Which mathematical formalism is suitable to describe the system we are interested in, and can better shed light on its functioning? Which computational tools do we need to simulate and analyse this system, and what kind of predictions the model is expected to give?”.

To answer these questions, in the first part of this seminar we will provide a brief overview of the most popular modeling approaches - ranging from static and qualitative models, to dynamic and quantitative models (e.g., interaction-based, constraint-based, logic-based and mechanism-based models) - highlighting the limits and the strengths of each strategy.

Some general tips will be given to select the most adequate modeling approach according to: 1) the biological questions that motivate the development of the model; 2) the size of the biological system; 3) the desired level of detail for the system formalization; 4) the availability of experimental data; 5) the running time of the computational analyses that the model will require.

The seminar will then focus on reaction-based modeling, a mechanism-based approach especially useful to describe and analyze networks of biochemical reactions. The seminar will provide some examples of reaction-based models, as well as the numerous advantages of this framework, such as its high level of detail and easy understandability, or the possibility to run both deterministic and stochastic simulations by exploiting the same model.

The seminar will end with a hands-on session on reaction-based modeling of simple biological systems. No particular notions in Mathematics or Computer Science are needed to actively participate in the hands-on session: only curiosity and the willingness to put oneself to the test are required!

### Graphics Processing Units for Life Sciences

Nowadays, mathematical modeling and computer simulation play a relevant role in the investigation of biological systems. However, in silico methods can represent serious computational challenges, which could be partially mitigated by the use of high-performance architectures. Among the possible choices, Graphics Processing Units (GPUs) present many advantages with respect to computer clusters, grid and cloud computing: they are affordable and power-efficient, giving access to tera-scale performances on common workstations. In this seminar, we will describe the most widespread library for general-purpose GPU computing – Nvidia CUDA – and we will summarize the results of GPUs applied to Systems Biology, Computational Biology, and Bioinformatics. In particular, the state-of-the-art of the existing GPU-powered tools will be presented, focusing on the acceleration of deterministic, stochastic, and spatial simulations of complex biochemical systems.

### GPUs for Systems Biology, in practice

The second part of the seminar will be dedicated to a hands-on session, in which we will describe three examples of typical methods exploited in Systems Biology: parameter sweep, sensitivity analysis, and parameter estimation. We will implement, and discuss, three tools for the aforementioned tasks, all integrating the GPU simulators. All tools will be implemented using the Python language. In order to run the examples, the following libraries need to be installed: numpy, scipy, SAlib, matplotlib.

Finally, the seminar will conclude with a discussion about the future develoments in the field of general-purpose GPU computing for life sciences.

**Daniela Besozzi** received the Ph.D. in Computer Science (2004)
from the University of Milano, Italy. From December 2002 to September 2015 she has
been Assistant Professor at the University of Milano, since October 2015 she is
Associate Professor at the University of Milano-Bicocca, Italy. She is member of
the SYSBIO.IT Centre of Systems Biology, Italy.

Her research interests concern topics in Mathematics, Computer Science and Biology, with a particular attention to:

- the definition, simulation and analysis of mechanism-based and logic-based models of various biological systems (e.g., signal transduction pathways involved in genome stability and in cell growth; metabolic and signaling pathways involved in the regulation of programmed cell death in cancer; proliferation of cell populations in acute myeloid leukemia, etc.);
- the development of computational methods for parameter estimation and for the reverse engineering of biochemical reaction networks, exploiting various evolutionary computation methods (Particle Swarm Optimization, Cartesian Genetic Programming, Evolutionary Petri Nets, etc.);
- the development of self-tuning evolutionary computation methods for the optimization of complex problems, by exploiting fuzzy rule-based systems;
- the development of computational intelligence methods for the prediction of 3D molecular structures;
- the exploitation of general-purpose GPU computing in Computational Systems Biology.

**Marco S. Nobile** received his Ph.D. in computer science from the
University of Milano-Bicocca, Italy, in 2015. He is currently a Research Fellow at
the Department of Informatics, Systems and Communication at the University of
Milano-Bicocca, Italy. He is also member of the SYSBIO.IT Centre of Systems Biology,
Italy. His research interests include high-performance computing, GPGPU computing,
evolutionary computation, swarm intelligence, bio-inspired algorithms, computational
biology, systems biology and synthetic biology.