The development of high throughput technologies have revolutionized numerous experimental approaches in the life sciences. The huge amount of data generated by these techniques need the development of new algorithms, computational tools and user-friendly interfaces that facilitates data analysis. The bioinformatics and computational biology area of the Center will develop the bioinformatics structure for analyzing data produced by next generation sequencing using high performance computing. The main projects include genome sequence assembly, transcriptome and SNPs studies and metagenomic and metatranscriptomic analysis.
CCES constitutes a unified venue of excellence, the first of its kind in Brazil, in which the current expertise of its participants in the areas of high-performance computer simulations, modeling, and data-intensive computing in general, can be synergistically multiplied by the proximity among the participants. We expect to be able to address and solve a variety of problems at the forefront of Science, including the areas of Computational Mechanical Engineering (viscoelastic materials, particulate and semi-continuum problems, boundary element methods and analysis), Computational Geophysics (e.g., seismic image processing and construction, signal analysis for deep water oil exploration), Computational Physics and Chemistry (quantum and classical atomistic simulations of nanomaterials, electronic structure of molecules and solids, structure and dynamics of proteins and other biomolecular systems, especially devoted to bioenergy and human health problems), Computational Biology and Bioinformatics (DNA sequence analysis and assembly, comparative genomics, protein structure prediction, high-throughput image analysis and drug discovery, systems biology), and Computer Science (high performance computing mechanisms and code parallelization, scientific data management, workflow management systems, domain specific languages, and various aspects of eScience). The Center will provide an efficient means of promoting integration among computational scientists of different expertise, enabling them to conduct high-level research in new ways, by exploiting complementary knowledge and research practices. In consonance with the emerging field of eScience, broad impacts are expected on ways to cope with advanced techniques in parallelism extraction and multi-core architectures, as well as processing, curation, and provenance of large volumes of information and data.
- Velocity analysis and velocity model building;
- Seismic tomography;
- Time and depth migration;
- Time-to-depth conversion;
- True amplitudes;
- Detection and processing of diffractions
- New or non-conventional signal processing;
- Compressive sensing;
- Vertical, lateral and super resolutions.
The analysis and proper description of material behavior in multi-scales, ranging from molecular behavior to the continuum approach have become a regular engineering task. Multi-scale and multi-domain problems have extended enormously the range of problems that engineers are tackling. Computational mechanical engineers at the Center have been developing and writing FEM, BEM and DEM codes and other numerical schemes to describe a significant list of engineering problems, including:
- the dynamic response of multi-domain systems, such as soil-structure interaction, fluid-structure interaction, soil-fluid-structure interaction;
- wave propagation in fluids, elastic, viscoelastic and poro-elastic materials;
- vibration analysis of nano-facilities;
- problems at the interface between the biosphere and the fossil oil reservoirs in offshore operations.
In the area of quantum chemical electronic structure problems, researchers at the Center will combine density functional theory (DFT) and ab initio calculations at different levels of theory in order to obtain accurate properties of atoms and molecules that have been used successfully in substitution of more sophisticated and computationally expensive higher level methods.
The infrastructure will focus on data sharing, data linking, ontology and taxonomy management, all of which are acknowledged to pose open research problems. More specifically, the research needed to create this infrastructure will concentrate on three directions:
- semantic web aspects, involving ontology creation and management, and linked data;
- spatio-temporal data management and database versioning mechanisms to help multiscale data management;
- scientific workflow specification and execution mechanisms.
The solution of modern scientific problems require efficient modeling techniques, high-performance computing and the ability to extract knowledge from a massive amount of data. In order to address these issues the Center focuses in three major research themes:
- designing domain specific languages, tools and workflows which enable scientists to easily model simulation experiments and inspect their results, by hiding most of the complexity associated with the involved computational and programming issues;
- working together with other areas to address problems that can benefit from the advanced parallelism extraction techniques and multi-core and GPU architectures;
- identify scientific problems in Physics, Molecular Biology and Chemistry which could be better understood via an algorithmic modeling approach.
Proper descriptions of the physical and chemical properties of carbon nanomaterials require both large scale classical and hybrid quantum-classical simulations which can only be carried out in realistic systems containing millions of atoms taking advantage of the advent of GPU-based computations. At the Center we gather the necessary conditions to develop, test, and apply these new technologies to the fully atomistic description of a large family of nanostructures. In particular, we intend to investigate carbon nanotube forest and yarns, large-area graphene and graphyne, graphene-oxide sensors, graphene/graphyne-based hydrogen storage materials, and metal-organic frameworks (MOFs).
Researchers at the Center seek solutions in frontier problems in quantum and classical molecular simulations to address protein-related human diseases, bioenergy, and carbon nanomaterials. The main systems that we are currently working on include:
- Proteins associated to various types of cancer in humans (e.g., prostate and breast), as well as metabolic disorders, such as diabetes, obesity, and inflammation;
- Human variants of hemoglobins associated with several sickle-cell pathologies in Brazilian populations;
- Molecular simulations of carbohydrate modifying enzymes for the production of renewable biofuels and chemical commodities from lignocellulosic biomass;
- Molecular aspects of plant cell wall architecture;
- Molecular understanding of structure/function relations in photosynthetic protein complexes.