hidden COVID-19 Updates

Please visit the UC San Diego Return to Learn page for up-to-date campus COVID-19 guidelines.

News Release

Natural Product Discovery Tool Wins UC San Diego Center for Microbiome Innovation (CMI) Grand Challenges Award

Pavel Pevzner and Bahar Behsaz
Bahar Behsaz and Pavel Pevzner.
Photo credit: Erik Jepson, UC San Diego

La Jolla, Calif., September 26, 2018—Each year, the Center for Microbiome Innovation (CMI), based on feedback from their industry partners (BASF, IBM, Illumina, Janssen, Metagenics, Panasas, Pfizer, and now Sanitarium) regarding current major challenges that could benefit from additional resources, provide $200,000 in funding towards a Grand Challenge award. The theme of this edition was about integration on multi-omics methods to enable an increased discovery rate.

This year, the development of a new computational tool from computer science and engineering professor and CMI faculty member Pavel Pevzner designed to address challenges in natural product discovery has been selected as the winner of the 2018 UC San Diego Center for Microbiome Innovation (CMI) Grand Challenges Award by the industry panel based on its strengths in terms of feasibility, innovation, collaboration,  and potential commercial impact.

The tool, named CycloMiner, is a pipeline for discovering and refining potential therapeutic cyclic non-ribosomal peptides (CNRPs)—a biomedically important class of natural products that includes such antibiotics as vancomycin and daptomycin, which are used extensively to treat bacterial infections of the skin and underlying tissues, and infections that have entered the bloodstream.

“Many antibiotics are natural products: compounds produced by microorganisms to give them an evolutionary advantage,” said Bahar Behsaz, a CMI graduate student fellow in Pevzner’s lab who will lead the project. “A lot of them have cyclic backbones, which give them several favorable properties as drugs. Unfortunately, that also makes them very difficult to find.”

The objective of CycloMiner is to significantly reduce the computational complexity required to identify promising drug candidates in this class of natural products and automate the process for high-throughput cyclopeptide discovery. 

In line with the mandate for this year award, the tool works by integrating cross-analysis of the genomes of cyclopeptide-producing bacteria and mass spectra of cyclopeptides these bacteria synthesize.

Pavel Pevzner and Bahar Behsaz
Pavel Pevzner and Bahar Behsaz. 
Photo credit: Erik Jepson, UC San Diego

The Need for a New Tool

For decades, the pace of natural drug discovery slowed; it has only been in recent years that new technologies have kicked it back into gear. Advances in mass spectrometry (MS), as well as genomics and metagenomics technologies, have enabled high-throughput data generation while bioinformatics tools are enabling analysis—but they are charged with keeping pace with the technological advances.

One such tool, already developed by UC San Diego scientists, is the Global Natural Products Social Molecular Networking (GNPS) infrastructure launched in 2016 by UC San Diego Skaggs School of Pharmacy and Pharmaceutical Sciences professors Nuno Bandeira (also a professor of computer science and engineering) and Pieter Dorrestein. GNPS, based on the concept of “molecular networks” (proposed by Bandeira and Pevzner in 2007), already houses over one billion mass spectra of natural products. But even though it is an untapped resource for the discovery of new antibiotics, one limitation for this tool is that only a small fraction of the deposited spectra are currently annotated.

While the annotation rate is steadily increasing, it can be expedited by a tool such as Cyclominer. “Using genome mining tools, we can generate millions of conjectures about the unknown cyclopeptide sequence,” said Behsaz. “But, it is unclear which one is correct. Using mass spectrometry tools, we can generate millions of different conjectures about the sequence of the same cyclopeptide but again, it is unclear which one is correct. CycloMiner aims to solve the puzzle of matching up the genomic and spectral data to figure out which of these conjectures is correct.”

This integration of genome mining and mass spectra represents a new peptidogenomics approach—a term coined by Pieter Dorrestein.

Once CycloMiner identifies candidates, it further leverages database search tools to provide the annotations, leaving the users with everything they need to isolate the compound. From there, it’s on to investigation.

Potential Commercial Impact

Cyclopeptides represent a multibillion-dollar market in antibiotics, immunosuppressants and anti-tumor agents, and include widely used drugs like cyclosporine, vancomycin, daptomycin, and many others. Future cyclopeptide discoveries facilitated by CycloMiner will provide a new resource for automated drug discovery with potential commercial value. 

“Cyclopeptides are of high interest to biomedical and agricultural research where biocontrol of microbes is desired,” Kirk Francis, Manager for Trait Knowledge & Performance Biologicals at BASF Bioscience, which is one of CMI’s industry partners who reviewed the proposals. “This is an innovative approach to mine existing tools, collections, and data sets to bring benefits to the entire bioscience community.”

Behsaz anticipates that CycloMiner will generate a lot of compounds that can be further tested for bioactivity. “In fact, we’ve already seen that with our pilot version. The $200,000 award from CMI will assist us with the implementation and integration of CycloMiner into the GNPS user interface.”

All of the tools in the CycloMiner pipeline (including CycloMiner itself) are open-source and will be available for licensing after completion of the project. 

This project brings together multiple omics approaches (including metagenomics, transcriptomics, peptidogenomics, and metabolomics) and extends the CMI and Pevzner’s existing collaborations with industry and academia. 

Other collaborators on this project and at the interface of bacterial genomics and natural products discovery include Anand Patel (Digital Proteomics), Pieter Dorrestein (UC San Diego Skaggs School of Pharmacy and Pharmaceutical Sciences), Chris Dupont and Anna Edlund (J. Craig Venter Institute), Bill Gerwick (Scripps Institution of Oceanography), Rob Knight (UC San Diego Jacobs School of Engineering and School of Medicine), Hosein Mohimani (Carnegie Mellon University), Marnix Medema (Wageningen University) and Andrey Osterman (Sanford Burnham Prebys Medical Discovery Institute).

Media Contacts

Erin Bateman
Center for Microbiome Innovation