Main Focus
"In silico structural biology correlated with experiment"
Mission Statement
The Department of Bioinformatics & Structural Biochemistry (DBSB) was set in 1999 aiming to consistently implement computational biology techniques - bioinformatics, modeling, simulation - and use them to guide experimental research in molecular biology and biochemistry.
Overview
Historically the main focus of DBSB was the investigation of glycoprotein (GP) folding & degradation and the relation between glycosylation and GP structure. On this line DBSB works in co-ordination with the Department of Molecular Cell Biology and colleagues from the Department of Biochemistry, University of Oxford, UK. Results have shed light onto the structure of glycans attached to the nascent protein chain and their role in the ER quality control and degradation [1,7-5,14,20,33]. The way glycosylation affects GP structure was also assessed using bioinformatics approach. To this end DBSB has developed SAGS: "Structural Assessment of Glycosylation Sites" - a comprehensive Data Base curating structural information on glycans and glycoproteins, with significant applications in modeling biological systems at molecular level [4,13,15,20-21,32].
DBSB was also involved in more general aspects of the physics of protein folding and structure. Configurations along the folding pathway and in the native state are degenerate and the experimental data has to be interpreted in the frame of statistical mechanics. In collaboration with colleagues from CEA Saclay and Heidelberg University we have investigated unfolded and native protein states using molecular simulation in combination with Small Angle and Quasielastic Neutron Scattering.
These studies have resulted in a better understanding of the folding process and the motility gradients in proteins explaining their stability and ability to interact with other molecules [2,3,8-12].
Combining bioinformatics with principles of protein folding and dynamics results in many applications in molecular life sciences such as building probabilistic models when the experimental structures are not available, or probing in silico biomolecular interactions.
Presently a significant part of our work is devoted to developing techniques in this field and applying them to a variety of problems in structural biology and molecular medicine. For example in collaboration with colleagues from IBAR and Oxford University we continue to investigate the structural aspects of processes along the secretory pathway; with colleagues from Yale School of Medicine USA; Levin Cancer Institute USA; Warsaw University Poland we are looking into protein-DNA interactions in a number of systems relevant in molecular medicine & biotechnology[29,34,36,38,42,43]; with groups from Universities of California - Berkeley and Davis, Wageningen, Haifa, Zurich, MPI-Koln, INRA-France etc we are involved in investigating the molecular basis of plant-pathogen interaction [16-17,22-24,26-27,31,39,41,45-47].
The models generated so far in this process allowed us to make predictions upon systems behavior. Subsequently predictions were successfully validated experimentally and resulted in a better understanding of the molecular bases underlying the investigated biological functions - some of them with potentially important applications in molecular medicine, biotechnology and pharmacology [16-52].
Along the years, DBSB work was financed by a large number of national and international grants among of which we mention 1 EU-IP-FP6 ("Bioexploit": 2005-2011); 1 EU-FP5 ("Nonema": 2001-2004); 2 UK - Wellcome Trust ( "Tyrosinase Folding": 1998-2001; "Glycoprotein Database": 2002-2005) or the national grant PN-II-ID-PCE 168 ("in silico": 2007-2011); PN2-ID-PCE-2011-3-0342 "Modeling molecular complexes and assemblies with experimental and bioinformatic constraints" (2011-2016); L.Spiridon - PN-III-P1-1.1.-TE-2016-1852 "Free energy prediction of biomolecular processes using high speed robotic algorithms" (2018-2020); PN-III-P4-ID-PCE-2020-2444 "Usage of high speed robotic algorithms for in-silico assisted experimental research of large biomolecular systems" (2021-2023), and many other national research consortiums.
DBBS was also member of the MC of COST Action "Sustain" FA1208-11941: "Pathogen-informed strategies for sustainable broad-spectrum crop resistance" (2013-2018) while AJ-P was also Co-PI in the USA-NIH grant 4R37 AI032524 "Structure of RAG1-RAG2-DNA complexes" (2012-2017) and NIH 1R01 AI137079-01A1/2018 "Function and evolutionary origin of RAG1 endonuclease" carried by Professor David Schatz from the Departmentof of Immunobiology, Yale School of Medicine.
Currently our group is involved in carying out two major FEDR funded consortium projects of national interest: ROGEN - Developing Genomic Research in Romania (2024 - 2029) and CANTAVAC - Development of Translational Research for Vaccines, Serums and other Biological Medicines (2024 - 2029) as well as in the EU Horizon Project COMBINE HLTH-2024-DISEASE-08
Our group continues to develop research strategies in computational driven experimental research by currently focussing on developing bioinformatics resources, new machine learning and statistical mechanics techniques for structural prediction and a new generation of highly effective molecular simulation methods.
For instance Laurentiu Spiridon who returned at the fall of 2016 from PostDoc stages at Illinois Institute of Thechnology is currently developping extremely efficient, robotic based algorithms for HMC simulations.
Teodor Sulea, Alexandru Buceag and Anca Iacob - PhD students are using MD simulation techniques and also tesing the new HMC algorithms in modeling structures, processes and interactions various biomolecular systems while Eliza Martin, PostDoc has developed Machine Learning techniques for fast identification of the main CC/TIR; NBS and LRR domain signatures, critical for the delineation of NOD like receptors in large sequence datasests. This work was instrumental in generating NLRscape - the Atlas of plant NOD like receptors (NLRs) - a web resource developed by DBSB for plant research community interested in NLR evolutionary dynamics and artificial evolution attempts aimed at expanding NLR diversity via bioengineering.
Another priority of DBSB, in which Cristian Munteanu plays the key role is now to coupling computational techniques with Mass Spectrometry, Surface Plasmon Resonance and data derived from the Highthroughput Drug Screening Platform of IBAR, aiming to step up the scale of biological system investigation to global proteome and interactome level [35,40, 44].
This project aims to address a number of structural aspects related to key elements of the plant immune system and its pathogen interactors using a combined approach intricating experimental and computational steps. To this end we intend to build on our previous results in the field and further develop experimental, bioinformatics and molecular modeling methods appropriate for solving the specific problems implied by this proposal.
By developing new methods which leverage high-speed robotics algorithms, we undertake the effort to enhance free energy prediction for biomolecular processes. Our focus includes integrating these advancements into existing APIs and applying them to key areas of molecular biosciences such as plant-pathogen interaction, drug design, and immunotherapy.
This project aims to enhance our understanding of the complex biochemical processes involved in Alzheimer's disease and overcome existing limitations in amyloid research by developing an analytical system capable of detecting amyloid concentration during the aggregation stage and advancing drug development in this field.
This project aims to improve protein engineering for medical and food security applications by leveraging advanced in silico assistance. Our software, Robosample, combines robotics algorithms with Gibbs sampling to accurately recover biomolecular free energy surfaces. We will further integrate robotics and molecular simulation by introducing new non-equilibrium sampling methods and novel robotic joints.
DBBS developed Databases:
1. SAGS DB - Structural Assesment of Glycosylation Site Database
2. NLRscape - an Atlas of plant NOD like receptors
DBBS developed Software:
1. Robosample - a Constraint Diynamics HMC (CDHMC) Molecular Simulation Program
1. NLRexpress - a ML program for swift detection of Nod Like Receptors: NLR/CNL/TNL
Simulation Hardware:
12 Graphic Stations (overall power ~250 TFlops, storage 70 TB)
1 HP-HPC Cluster (4 TFlop - effective power);
1 Bull-HPC Cluster
Simulation Software: Charmm, Amber, NAMD
Modeling Hardware: 1 HP-Graphic Station ProLiant WS460c; 1 Octane 2x600 SGI Workstation
Modeling Software: Accelrys Discovery Studio, BCI Raptor, Modeller etc