Research

I relocated to Singapore in 2017 to take a position as a Research Fellow at the National University of Singapore (NUS), where I worked at the Cancer Science Institute (CSI), under the supervision of Prof. Daniel G. Tenen. In this second postdoctoral position, I contributed novel bioinformatics approaches in multiple multidisciplinary projects to try to uncover the molecular regulatory mechanisms at the RNA level and their implication in the onset and progression of cancer. This mechanistic understanding of cancer pathology will be key to predict candidate molecular targets and rational intervention strategies for reverting disease phenotypes. More specifically, I was involved in a collaboration with the Hematology/Oncology Department, at the Beth Israel Deaconess Medical Center, Harvard Medical School, for the profiling of patient primary cells for the identification of the regulatory dysfunctions triggering leukemic transformation resulting in the development of Acute Myeloid Leukemia (AML). In this project, we studied the variation of the DNA methylation profiles and the interacting networks between RNAs and DNA methyltransferase 1 (RNA-DNMT1 interaction networks), and how these changes at the epigenetic regulatory level trigger the phenotypical changes associated with disease progression from Myelodysplastic Syndrome (MDS) (i.e., MDS ⇒ AML), and in Remission (REM) from AML after treatment (i.e., AML ⇒ REM). As a result of this project, we provided a mechanistic understanding of the interplay between the epigenetics and gene expression levels during AML transformation, and the different dysregulations arising among patients. (As part of this collaboration project I spent a 5-month Research Stay in the Hematology/Oncology Department, at the Beth Israel Deaconess Medical Center, Harvard University)

After obtaining my Ph.D. in Biochemistry in April 2014 at the University of Zaragoza, I moved to Luxembourg to join the Group of Computational Biology, at the Luxembourg Centre for Systems Biomedicine of the University of Luxembourg as a Research Associate under the supervision of Assoc. Prof. Antonio del Sol. During these three intense years I had the opportunity to participate in many thrilling and exciting projects for the application of network models to study phenotypic transitions in complex biological processes, such as cellular reprogramming or differentiation, and disease onset and progression.

• Systems Biology study of complex cellular phenotypic transitions

In these projects, leveraging from high-throughput genomics data, we built comprehensive genome-wide gene regulatory networks models underlying the gene expression patterns of different cellular states (i.e., pluripotent ⇒ differentiated, or healthy ⇒ disease) and used them to identify the network regulatory determinants stabilising such phenotypes. The identification of these network motifs defining the phenotype-specific network attractors (i.e., combination of network circuits defining the meta-stable network state characterising a given cellular phenotype) allowed us to perform accurate prediction of candidate genes whose perturbations could trigger a desired phenotypic transition, such as the identification of the lineage specifiers triggering differentiation from a pluripotent cellular state to a specific cell type. We tested this computational approach in many different case studies, including Hematopoiesis, and the prediction of disease-gene-drug relationships for the identification of candidate genes and chemicals for reverting disease onset and progression.

In 2007 I moved to Spain to complete my Master and Ph.D. in Biochemistry at the University of Zaragoza in the DeveProtMol Group, under the supervision of Professor Javier Sancho. As part of my postgraduate projects, I started working on bioinformatics strategies to study protein folding and conformational diseases, using Molecular Dynamics. A detailed description of the projects follows:

• Bioinformatics prediction of unstable regions in proteins

Local flexibility is an intrinsic property of polypeptide chains, with critical implications in the structure and function of proteins. This inherent dynamics includes local fluctuations of flexible loops or active-site residues, concerted motions of entire subdomains, and large-scale rearrangements involving partial unfolding of the native state. These structural transitions are mediated by intermediate species that are central in protein folding routes. We have developed a methodology to predict local unstable regions of proteins based on the physicochemical and geometric characteristics of buried protein interfaces. Our methodology is based on the hypothesis that proteins contain conserved buried interfaces of high polarity and low packing density, which make them unstable and are involved in local unfolding events.

As part of this research line, we have also established a collaborative project with the group of Prof. Salvador Ventura, at the Institut de Biotecnologia i de Biomedicina from the Universitat Autònoma de Barcelona, to develop computational methodologies to predict prion forming domains in the complete proteomes of organisms. This study has resulted in the largest set of prion proteins reported so far, in which we have found the presence of putative prion proteins in all taxa, from viruses and archaea to plants and higher eukaryotes, and found that most organisms encode evolutionarily unrelated proteins with susceptibility to behave as prions. This kind of computational methodologies could be of great importance to identify potential targets for further experimental testing and to try to reach a deeper understanding of prions’ functional and regulatory mechanisms. We also have an ongoing project for the development of a publicly available database of putative prion domains in the complete proteomes of organisms PrionScan, on which we make all our predictions freely accessible to the scientific community, and also including a bundle to our model to allow researchers to process their own sequences in the search for prion-like domains.

• Predicting abnormal phenotypes caused by SNPs in Conformational Diseases

Mutations in the LDL receptor cause Familial Hypercholesterolemia (FH), a genetic disorder that constitutes a worldwide health problem. We have developed computational strategies to anticipate the impact of SNP in the stability, structure, and functionality of the LR5 module of the Low-Density Lipoprotein Receptor (LDL-r) using short-range molecular dynamics (MD). In this approach, we combine different computational strategies to analyse the dynamical evolution of all possible single-point mutants to explore the extent of the structural changes they could cause in the tertiary structure of the module. These analyses have given us the opportunity to follow the evolution of the trajectories and characterise mutations that could cause significant structural distortions and might be related to disease-like phenotypes. This study will be useful to establish strategies to make an anticipated computational diagnosis of FH phenotypes and to study the structural implication of SNPs in the development of the disease. This in turn may also prove the viability of performing computational anticipated diagnosis of human conformational diseases on an even larger scale. (During the completion of this study I had the great opportunity to spend 3-months in a research stay at the group of Prof. Modesto Orozco, at the Institute for Research in Biomedicine, Universitat de Barcelona)

In the past years, I've been linked to various research projects which combine multiple areas of Bioinformatics and Molecular Biology. These include the study of bacterial transcriptional regulatory networks from a topological, functional and evolutionary point of view. During this research, I've had the great pleasure to work under the supervision of, or in collaboration with, talented people from different countries, and to spend time in research fellowships in institutions from Brazil and Mexico. A detailed description of this research line follows:

• Prokaryotic transcriptional networks, transcription factors and the evolution of regulatory circuitry

Starting from the knowledge of TF-cognate site links in E. coli we developed a strategy to predict a subset of the transcriptional regulatory networks in various organisms from the class gamma-proteobacteria. This work gave us the opportunity to develop the database TRACTOR, which is publicly available and encompasses a worth of information regarding the architecture of the regulatory networks of these microorganisms, most of which are important human and animal pathogens. Using this information we also conducted studies of the impact of TF-DNA interactions in determining the evolution of regulatory networks and in deciding the rewiring of the networks during evolution. We've also developed novel structural methodologies to predict new TF-binding sites in complete genomes, which have been used to build knowledge-base databases such as 3dfootprint, a website for the study of the structural characteristics of protein-DNA interactions.

Collaborators:

• Dr. Abel González Pérez, now at the Group of Biomedical Genomics of the Institute for Research in Biomedicine, Universitat de Barcelona, Barcelona, España

• Dr. Ana Tereza Ribeiro de Vasconçelos, head of the LabInfo Group at the National Laboratory for Scientific Computing, Petrópolis-Rio de Janeiro, Brazil

• Dr. Julio Collado-Vides, head of the Computational Genomics Group at the Center of Genomic Sciences from the UNAM, Cuernavaca, Mexico

• Dr. Bruno Contreras-Moreira, head of the Laboratory of Computational Biology from Aula Dei Experimental Station-CSIC, Zaragoza, España