Chemistry matters. Join us to get the news you need.

If you have an ACS member number, please enter it here so we can link this account to your membership. (optional)

ACS values your privacy. By submitting your information, you are gaining access to C&EN and subscribing to our weekly newsletter. We use the information you provide to make your reading experience better, and we will never sell your data to third party members.


Green Chemistry

App predicts waste created during syntheses

Web tool could make it easier for pharma chemists to pick green synthetic routes

by Sam Lemonick
January 24, 2019 | APPEARED IN VOLUME 97, ISSUE 4


As you make dinner and clean, chop, and measure out your ingredients, little piles of waste grow on the counter or get ferried to the trashcan or compost bin. Pharmaceutical companies produce their own waste, sometimes in the form of solvents and unwanted byproducts, as they synthesize drug molecules.

To help pharma chemists find less wasteful routes to their products, members of the American Chemical Society’s Green Chemistry Institute have developed a web-based tool to predict the efficiencies of syntheses (ChemRxiv 2019, DOI: 10.26434/chemrxiv.7594646.v1). ACS publishes C&EN.

Credit: CEN

The group, which included representatives from 10 major pharma companies, collected data from those companies’ processes and the literature to build a database of about 2,000 reactions commonly used to make drugs. They used a metric called process mass intensity (PMI)—the ratio of the mass of materials used in a reaction to the mass of the products—as a measure of the efficiency of each reaction.

When users input a proposed synthetic route, the app calculates the estimated PMI of each step and the overall process, including uncertainties for the values. For instance, the tool predicted a PMI of 1,202 for a process to make Bristol-Myers Squibb’s drug brivanib alaninate. The actual PMI was 1,488, which fell within the app’s 95% confidence interval. PMIs for six of the 15 real-world processes tested by the researchers fell within the app’s 95% confidence interval.

“[The tool’s] chief value lies not in the exact value of any one prediction made, but in the ease with which it allows chemists to compare potential routes,” says Helen Sneddon, a scientific director specializing in green chemistry at GlaxoSmithKline and a member of the ACS GCI who was not an author of the paper. Juan Colberg, a green chemistry project leader at Pfizer who works with one of the paper’s authors, says the app could save chemists time by identifying which routes are likely to be greenest before they test them in the lab.

The authors declined to speak with C&EN except to confirm the scientific accuracy of our reporting to comply with the embargo policies of the journal where they have submitted their manuscript.



This article has been sent to the following recipient:

john andraos (February 19, 2019 9:49 AM)
I read with interest your news announcement of an App that “predicts pharma process waste” which is already available from the ACS GCI website ( Readers of this magazine should be informed that unfortunately this App does NOT yield correct results for the process mass intensity (PMI) for actual or conjectured synthesis plans as advertised. Specifically, the App yields results that significantly underestimate the actual global PMIs for experimental or hypothesized synthesis plans by as much as 50 % depending on inputted reaction yields, because it uses a flawed methodology which calculates global E-factors by naively summing step E-factors for a synthesis plan. PMIs are obtained directly from E-factors according to PMI = E + 1. The lower the step reaction yields are, the higher is the discrepancy between the calculated App result and the true value of PMI. Moreover, the results of the App unfortunately do not agree with other published work on the same subject by other scientists who developed web-based and software tools such as Environmental Assessment Tool for Organic Synthesis (EATOS) developed by Marco Eissen in 2001, Jacques Augé’s algorithm developed in 2008, and my own algorithm developed since 2005 which uses ubiquitously available Microsoft Excel spreadsheets that everyone has on their desktop. Amazingly, these prior works were conspicuously absent from the report as key precursors that laid the foundation for the development of the App. Interested readers who wish to practice green chemistry principles in synthesis design in a reliable way should consult the latest streamlined algorithm available at Green Processing and Synthesis 2019, doi: 10.1515/gps-2018-0131 which calculates step, cumulative, and overall PMIs and E-factors for any synthesis plan regardless of complexity using a very simple Excel spreadsheet. In that work the algorithm was rigorously tested on various literature syntheses of the API apixaban. Also, one can find a mathematical proof that the global E-factor for a synthesis plan is NOT the sum of the reaction step E-factors as used by the App. This latest peer reviewed paper also appeared on ChemRxiv (doi: 10.26434/chemrxiv.6686729.v1) in June 2018 well before the one cited in the C&EN announcement but was also not mentioned in the same report. I point all this out because your piece ignored prior work and endorsed something that had not passed the gauntlet of peer review, in addition to the obvious computational flaws contained in the App. By hastily printing this announcement, including positive endorsement by a member of ACS GCI, you have essentially usurped the proper course of peer review. Reporters of C&EN would be well advised to be very careful in reporting announcements from ChemRxiv which have not passed the test of peer review, particularly when it is known that a given paper is undergoing that process simultaneously. ChemRxiv is currently being used by authors for two main reasons: (1) for scientists to claim territory in their area of research, especially when they are aware of possible competitors in the same field; or (2) as a dumping ground for papers that would not otherwise pass the peer review process because they are based on research that is deemed derivative and thus would not pass the litmus test of being “new and innovative”. In this case, in light of the hot topic subject matter, the authors of the present App chose the first option. C&EN would be well advised to change its policies in advertising work from ChemRxiv in light of the obvious miscue with this App announcement. I would suggest, to be on the safe side, that C&EN stay clear of advertising apparent “hot topic” items from this source altogether.

Leave A Comment

*Required to comment