Open access chemoinformatics to tackle antimicrobial resistance
As part of MedChemNet’s focus on antimicrobial drug discovery, I discuss the work being done by the Community for Open Antimicrobial Drug Discovery (CO-ADD; Australia) to tackle the growing problem of drug resistance through providing a free screening platform for chemists.
There has been a lot of discussion about big data and open science in the news recently. What role do these fields have in drug discovery?
Big data is playing an ever increasing role in the discovery of novel drugs, as the synthesis and screening of larger compound collections becomes more and more affordable and efficient. However, while large pharma companies have been collecting and utilizing large data sets for their drug discovery and development programs, the utilization in the academic environment is still underdeveloped. Especially chemoinformatics, in this context, the relationship between chemical structures and biological activity, has not reached its potential. There are several silos of big data, such as PubChem, ChEMBL, linked by chemical structures services, like ChemSpider or ChEMBL’s UniChem. In addition, increasing numbers of chemical structures from patents become available in online databases. Open PHACTS is a notable service, which is trying to provide a central platform to mine and analyse data from different sources. However, while the wealth of chemical structures in online databases increases, the link to, and network of biological data still remains underdeveloped. Compared to bioinformatics data, no submission of data to online databases before publication is required, leaving the collections of big data to the mining of publications. This limits the coverage and quality of the biological activity data linked to chemical structures, as negative results are not always published and descriptions of assay methods usually lack enough details to compare results from different labs. In addition, apparent commercial reason prevent the publication of a large proportion of data.
Open Access facilities, like CO-ADD, will improve the coverage and quality of big data. Apart from providing a platform for collaboration and opportunities to engage in new research fields, open access screening facilities are able to provide large consistent data sets, from consistent assays. More importantly, open access platforms are able to produce complete data sets, including negative results which contribute to our knowledge as much as positive results. On top of that, open access facilities are by definition open and not limited as much by IP restrictions. Big data and open access, in combination with a more unifying data system, has the potential to improve the drug discovery and development, especially in the academic environment.
Could you please describe the current research being carried out at CO-ADD?
The Community for Open Antimicrobial Drug Discovery (CO-ADD) provides an open and free screening platform, for testing compounds for activity against bacteria and fungi. The initiative invites chemists from around the world to send as little as 1 mg of their compounds, for screening against 5 ESKAPE pathogens (S. aureus, E. coli, P. aeruginosa, K. pneumonia and A. baumannii) and two fungal pathogens (C. albicans and C. neoformans). In case the compounds are active against one of the pathogens, further studies will be conduct with the compounds, confirming their activity in dose response assays as well as counter screens for cytotoxicity and haemolytic activity. In case the compounds display potential for further development we will extend the screening to a broader panel of pathogens, including multi-drug-resistant clinical isolates, and some mechanistical studies to evaluate potential mode of action or target. All screening data will be provided to the chemists without any restriction, as they can use the data for publication or application for further funding. While the screening is limited to academic groups only, we try to convince all types of chemists to participate in this initiative, especially if they have not been working in antimicrobial research, and have produced compounds for different purposes. The aim is to attract the broadest possible chemical diversity, as only a small proportion of this chemical diversity has actually been screened for antimicrobial activity. In addition, antimicrobials tend to defy some of the drug-like filters, which are usually applied to commercial screening libraries.
CO-ADD’s main aim, however, is to invigorate the antimicrobial research in the academic community, by providing free screening platform for any chemist and provide them with a data package to extend or, more importantly, start new antimicrobial research projects. CO-ADD is thereby able to act as collaborator in further drug developments, or as a hub to build extended collaborative networks. In addition, CO-ADD will assist the chemists in raising further funding, either through specific antimicrobial initiatives or through a CO-ADD internal funding model. This will give the chemists a great opportunity to add value to their chemical invention and property.
CO-ADD’s further aim is to significantly increase the public knowledge on antimicrobials by making the data publicly available. To achieve this, CO-ADD encourages the chemists to publish the data, either themselves or in collaboration with CO-ADD, and submit all the data including negative results to ChEMBL’s online data system. Chemists will be granted a grace period of 2 years, in which further funding applications, patent submissions or commercialisations can be achieved. CO-ADD aims to provide a large and free antimicrobial knowledge base, which can spawn new antimicrobial research.
What are the particular challenges faced in antimicrobial drug discovery?
Antimicrobial research faces currently significant challenges with a continuous decline in funding and people in this field. Most large pharma companies have existed antimicrobial drug developments in the last 10 years with only few companies left actively involved in the development of novel antimicrobials. One contributing factor is that the prices for antimicrobials are too low to make the drug development economical feasible. Just compare the ~$100 we pay for antibiotics to the >$1,000 we pay for anticancer drugs. A similar trend can be observed in the academic environment, with significant less grant money awarded to antimicrobial research than compared to anticancer research, resulting in a constant drain of experts from the field. A further scientific factor is the intrinsic ability of bacteria to develop resistance against drugs, potentially reducing the life span of drugs on the market. Without novel classes of antibiotics reaching the market, we are currently in the situation that for certain infections, like the ones caused by carbapenem-resistant Enterobacteriaceae (E.coli and K.pneumoniae ), we can only rely on a single antibiotic, colistin. And even so, patients have already been identified carrying colistin resistant bacteria. This led to the dire prediction that we are heading in to an era, in which simple antimicrobial infections might become untreatable, again. Government and Health institutions from many different countries have recognized the issue and are trying to reverse the trend by issuing special programs for the clinics; however, similar initiatives are required for the early drug discovery research as well.
Do you believe that big data will be increasingly valuable in the battle against antimicrobial resistance in the next 5-10 years?
Different approaches can be pursued in finding new and improved treatment for infection caused by multi-drug resistant microbes: develop specific treatment to reduce the bacterial drug-resistance, develop specific treatment to reduce the severity of the infections, and increase the arsenal of available antimicrobials. Together with improvements in rapid diagnosis and in clinical managment we should be able to contain the global health impact of antimicrobial resistance. And indeed health care providers are starting to implement big data tools for their strategic decisions, with further developments required to bring it to the front of the primary health care provider, and into the surgeries.
For drug discovery, big data will play an import role in making the development process of new antimicrobial more efficient. Extending our knowledge and making the knowledge publicly available has the potential to reduce the failure rate of drug discovery. Current drug development project relay mostly on ‘local’ experts within the collaboration network. Access to comprehensive big data would allow to make better design decisions, especially at earlier stages of the discovery process, where such expert networks might not have been established yet. The task is, to expand our knowledge with broader chemical diversity, broader activity data, including failures and negative data, build improved structure-activity relationships, and extend on potential off-target liabilities. In addition, the knowledge would need to be made more accessible for primary researchers, like medicinal chemists and microbiologists, which are the key drivers for the early discovery process.
What change, if any, would you like to see across the drug discovery community?
The drug discovery community is still relaying mainly on publications to distribute their findings, and even though online journals have improved the turnaround time for publishing results, a publication itself is not an efficient format to distribute scientific data. The community should establish efficient ways to collect the data into online systems and make them available for the global community. For example, making the submission to online data system a requirement for publication, similar to gene sequences or protein structures. In addition, there is a benefit to establish more open access facilities and public data frameworks, increasing the sharing, distribution and utilization of data, and thereby establishing a broader collaboration network. Great discoveries are still done through great collaborations, and their values do not only lay in their economic potential, but in their benefit to the community.
For more information on CO-ADD, please visit: www.co-add.org.
Information and links to public data bases: ChEMBL (www.ebi.ac.uk/chembl/), PubChem (https://pubchem.ncbi.nlm.nih.gov/), Open PHACTS (www.openphacts.org/), ChemSpider (www.chemspider.com/), ToxCast (www.epa.gov/chemical-research/toxicity-forecaster-toxcasttm-data)