Our first BMERC-sponsored Machine Learning Journal Club was a blast! 20+ students and researchers from across different disciplines and departments such as Biomedical Engineering and Chemistry attended the event. This journal club was led by Israel and Megan from Prof. Sandor Vajda's group. We learned about basic concepts of supervised learning, followed by a hands-on session using scikit-learn and Jupyter Notebook.
Boston University hosts high school students who are passionate about scientific research in the summer term. This is known as the Research in Science & Engineering (RISE) program. RISE students spend six weeks at the university and gain hands-on experience in the exciting cutting edge scientific research happening on campus. Research projects span a wide range of topics, including astronomy, biology, biomedical engineering, chemistry, electrical and computer engineering, mechanical engineering, medical laboratory research, neuroscience, physics, psychology, and public health.
This past summer BMERC affiliated labs accepted two RISE students, Raymond and Vijay. Raymond was mentored by graduate student Amanda Wakefield from the Vajda group. He conducted computational studies on protein binding sites. Vijay worked closely with Ashley Rebelo from Prof. Karen Allen's lab and worked on protein crystallization. At the end of the RISE program, both students presented their research at a poster symposium in front of an audience comprised of those familiar with research as well as the general public. Raymond's and Vijay's posters are titled "A Comparison of Computational and Experimental Methods for Determining Binding Sites of Proteins" and "Optimizing Crystallization Conditions of the N-terminal Domain of hAMPDA2 and Implementing a Baculovirus Expression System for hAMPD1 and 3" (see photos).
The ClusPro protein-protein server is the best performer in the CAPRI46-CASP13 prediction competition
ClusPro is a web server that performs rigid body docking of two proteins by sampling billions of conformations. The server has over 13,000 registered users and performs about 5,000 docking calculations each month. The two papers describing ClusPro have been cited over 1,500 times, and over 600 of these papers reported protein complex models built by the server. ClusPro has been participating in the CAPRI (Critical Assessment of Predicted Interactions), the ongoing communitywide experiment devoted to protein docking. In the CAPRI challenge, participating research groups and automated servers are given prediction targets, each being an unpublished experimentally determined structure of a protein-protein complex, and each group is expected predictions. The submitted models are evaluated by independent assessors. More recently CAPRI has been joined forces with the ongoing CASP (Critical Assessment of protein Structure Prediction), and most targets had to be predicted started from sequences rather than structures of component proteins.A homodimer prediction (cyan and blue) submitted by the ClusPro team which received a Medium quality score, overlapped with its crystal structure (wheat).
The most recent combined competition, CAPRI46-CASP13 was held during the summer of 2018 and has been evaluated on December 1-4 at the Iberostar Paraiso Maya resort on the Riviera Maya. The Vajda group was represented by graduate students Katie Porter and Israel Desta, who participated in the competition, in collaboration with the group of Prof. Dima Kozakov, a long-term member of BMERC, now an associate professor in the Department of Applied Mathematics at Stony Brook University. According to the evaluation at the meeting, ClusPro was the best performer in the server category. The server performance is described in the paper KA Porter et al., Template-Based Modeling by ClusPro in CASP13 and the Potential for Using Co-evolutionary Information in Docking, to be published in the journal Proteins: Structure, Function, and Bioinformatics.
Dr. Marcelo Castilho, a Professor at the Universidade Federal da Bahia, spent two months working with the Vajda lab during the summer of 2018. The first paper resulting from this collaboration, TQ Froes et al. Structure-based druggability assessment of anti-virulence targets from Pseudomonas Aeruginosa, was published on April 17, 2019, in the journal Current Protein & Peptide Science. Dr. Castilho is a very driven scientist, and we were happy to have him in the lab. He was also working on a second paper while in the Vajda lab, and the paper is close to getting finished.
The BMERC has accepted multiple high school students through the BU RISE program. These students work 40 hours a week for six weeks in one or more of our affiliated labs. This experience provides each of the students unparalleled access to both experimental and computational facilities.
One of the main area of our research, supported by the MIRA grant R35 GM118078, entitled Analysis and Prediction of Molecular Interactions, is investigating the interactions between proteins and small molecules, including drugs and metabolites. Bioactive small molecules, such as the products of cellular metabolism, natural products and synthetic organic compounds, are potent mediators of biological processes as ligands and allosteric regulators. Mapping their physical associations is therefore a critical but challenging task. To this end, we seek NIGMS administrative supplements to acquire a high performance Thermo Scientific Q-Exactive Plus Orbitrap Mass Spectrometer (QE+MS) to support our efforts to explore and exploit the dynamic interactions of small molecules with cellular proteins, biochemical pathways and signaling cascades. We plan to use this instrument together with our new colleague and collaborator, Andrew Emili, a recent recruit to Boston University and leader in using mass spectrometry to map protein interaction networks, to study protein-metabolite interactions (PMIs) in a rigorous experimental manner. Our computational modeling predicts many novel interactions that require stringent experimental validation, which currently stretches the limited MS capabilities available to us. In addition to providing valuable information on potential PMIs, mass spectrometry will address a major challenge in differentiating putative functional regulatory interactions from non-specific interactions that do not entail functionality. This problem is an excellent fit for the structural modeling methods developed in our lab, and we will thus work in a synergistic, iterative manner to apply computational tools to analyze and prioritize experimental interaction data provided by the proposed mass spectrometry instrumentation for the identification of PMIs that are most likely functional. Although establishing functionality will still require other biological tests (e.g. enzyme assays), we expect to develop systematic methods that reduce the number of compounds that need to be investigated in such a demanding, low throughput manner. Having access to a state-of-the-art protein-metabolite interaction platform and our collaborations with Dr. Emili will be very productive and lead to fruitful new avenues for our research program. The collaboration will also engage us with the team of Dr. Daniel Segre at Boston University, whose lab studies complex metabolic networks, and thus will directly contribute to the identification of novel protein-metabolite interactions. Dr. Segre’s research is supported by the grant R01GM121950 “A platform for mining, visualization and design of microbial interaction networks”, and he is also requesting an administrative supplement to support the acquisition of the proposed mass spectrometer. The University’s commitment is demonstrated by its investment in renovating the infrastructure for supporting this multi-user research instrument and by the Department of Biology’s provision of partial support for an ongoing service contract to ensure its operation.
The recently held BMERC open house featured posters from labs across the Chemistry and Biomedical Engineering departments. The event successfully allowed graduate students and faculty members to engage undergraduate students in conversations about the many research projects going on at BMERC. Many undergraduate students were surprised at the wide variety of labs and their projects at Boston University and expressed interest in pursuing graduate degrees. Check back for updates on upcoming poster sessions!
Zhuyezi Sun, a graduate student in the Vajda Lab, presents a poster on the ClusPro Server which was developed by the Vajda Lab.
Margarita Tararina, a graduate student in the Allen Lab, presents her work on determining the structure of a protein within the HAD-Superfamily.
Professor Vajda studies a poster on macrocycles authored by Lauren Viarengo (not pictured), a graduate student in the Whitty Lab. Israel Desta (left), a graduate student in the Vajda Lab, presents a poster about using machine learning to improve protein docking.
The Vajda Lab is excited to announce their new visiting fellow, Marcelo Santos Castilho, PhD. Marcelo received his MS in chemistry and his PhD in physics at the University of Sao Paulo. He currently holds a faculty position at the Federal University of Bahia.
This new paper describes our ClusPro web server, a widely used tool for protein-protein docking. ClusPro provides a simple interface for basic use, but it also offers a number of advanced options to modify the search. These include the removal of unstructured protein regions, application of attraction or repulsion, accounting for pairwise distance restraints, construction of homo-multimers, consideration of small-angle X-ray scattering (SAXS) data, and location of heparin-binding sites. Applications of ClusPro include docking X-ray or NMR structures of proteins, modeling antibody-antigen interactions, constructing the structure of multidomain proteins, building homo-oligomers, peptide docking, homology model docking, and more.