The computational biology laboratory of the Sapienza University in Rome is a multidisciplinary Unit established in 2001 composed by physicists, computer scientists, engineers, chemists, pharmaceutical chemists, biotechnologists and biologists. The head of the group, Anna Tramontano, has been working in the area of computational biology and bioinformatics (ante literam) since 1980 in the USA and Europe, in both academic and industrial setups.
Our group research activities are devoted to the computational analysis of biological data, to the development and improvement of methods for the analysis of genomes and proteomes and to the applications of methods to problems of biomedical interest. The group is involved in many world-wide initiatives aimed at analysing post-genomic biological data in order to improve our understanding of life at a molecular level.
Our recent research interests can be divided, with some overlap, into two main directions: development and/or improvements of computational biology methods and their application to the study of problems of biomedical interest.
Other relevant activities of the group concern the organization of the world-wide CASP (Critical Assessment of Techniques for Protein Structure Prediction) experiment and the participation to the Italian Epigenomics Flagship Project, where we are co-coordinating the Bioinformatics efforts.
NGS data analysis
The unit is interested in the analysis of the results derived by Next Generation Sequencing (NGS) experiments. In doing this we employ the commonly used strategies for the analysis of RNA-Seq (miRNA and mRNA) data as well analysis pipelines for ChIP-Seq. RNA-Seq analysis have the main aim of identifying genes differentially expressed in two or more samples. At the same time these analyses usually imply extending genome annotation, detecting splice variants and identification of new transcriptional units. We also pay attention to the molecular regulatory circuits involved in gene regulation. We developed an effective method to predict miRNA/target-genes.
The unit is involved in several projects, in collaboration with experimental groups.
Antibody structure and function
A finite number of antibodies must bind a virtually infinite range of foreign molecules and this process makes the immune system a major area for studies on molecular recognition. Basing on their sequence variability, immunoglobulin polypeptide chains are divided in two functionally distinct regions: the “variable region”, that is responsible for the binding of different antigens, and the conserved “constant region”, that is known to participate in signaling events, such as the activation of the complement cascade.
Our lab has contributed to the definition of rules relating sequence and structures of antibodies, developed methods for the prediction of their structure and maintains a database with integrated tools (DIGIT) and is now studying conformational adjustments that might play an important role in antigen recognition and in immunoglobulin function in general. We also developed a method to design humanized antibodies.
Study of pathogens
In the last years is we have exploited the knowledge of the Plasmodium and Schistosoma genomes and proteomes as well as our expertise in modelling, docking, and drug design, to develop novel inhibitors of key enzymes in both parasite lifecycles. Both parasites are blood feeders, employ similar methods to evade the immune system and are targets of an overlapping set of therapeutics. This creates both a problem, related to drug resistance, and an opportunity. The knowledge of a drug target in one parasite can lead to the identification of the unknown counterpart in the other. Similarly, an existing drug against one parasite can be used as lead for developing an inhibitor of the analogous target in the other.
In particular, we focused on molecules derived from chloroquine, arthemeter, oxadiazole-2-oxides and auranofin, known to be effective on both parasites, albeit with different activity and specificity and on the following candidate targets: SmTGR, PfTR, PfATP5, SmSERCA, PfHDP..
We also extende the concept of drug repositioning using information on validated targets of approved drugs. The combination of different state of the art techniques that we plan to employ in this project (homology detection, binding site structural comparison, docking) is expected to provide a lower number of false positive hits and therefore to reduce the number of compounds to be validated to an approachable figure.
Protein structure modeling
In this area we have developed a complete integrated pipeline for performing all steps of a modeling project (MODExplorer and MODAlign). This is an integrated tool aimed at exploring the sequence, structural and functional diversity in protein families useful in homology modeling and in analyzing protein families in general. It permits to manually modify the alignment, estimate the quality of the implied model while the alignment is being modified and, finally, to build a comparative model. We also developed method for ab initio prediction of protein structures based on an innovative use of replica exchange methods.
Protein interaction and peptide design
We have shown that, by automatically comparing and analyzing structurally similar regions of proteins of known structure interacting with a common partner, it is possible to identify mutually exclusive interactions present in the maps with a sensitivity of 70% and a specificity higher than 85% and that, in about three fourth of the correctly identified complexes, we also correctly recognize at least one residue belonging to the interaction interface.
We follow a general strategy based on the hypothesis that, if two proteins interact with the same region of a common protein and therefore their interactions cannot be simultaneous, they might share a common surface region mediating the interaction. To reliably identify these cases, we extract from a PPI all instances where at least three proteins of known structure interact with a common protein partner, compare their surface residues to identify structurally similar substructures comprising at least three residues and list the results together with the level of structural and sequence similarity of the matching residues. We have developed and made available a tool reproducing this strategy called Estrella.
There is a wide interest in designing peptides able to bind to a specific region of a protein with the aim of interfering with a known interaction or as starting point for the design of inhibitors. We developed PepComposer, a new tool for computational design of peptides binding to a given protein surface. PepComposer requires as input only the target protein structure and an approximate definition of the binding site. In our method, we first derive a set of peptide backbones scaffolds from monomeric proteins that harbor the same backbone arrangement as the binding site of the protein of interest. Next, we design optimal sequences for the identified peptide scaffolds. The method is fully automatic and available as a web server.