How Supply Chain Digitalization Efforts Will Help Increase the Efficiency of Drug Approvals
Reflections on Leveraging Chemistry Data in Machine Learning

Two decades of NMR Software—Through the Eyes of a Chemist Turned NMR Spectroscopist

by Sanji Bhal, Director, Marketing Communications

I recently had the pleasure of speaking with Norman Aubry—a long term user of ACD/Labs’ NMR software.  After 40 minutes discussing his career, our shared training as synthetic chemists, and his current situation (semi-retired but working from home as a specialist in structure characterization by NMR) I thought others would also enjoy hearing his story.

Naubry croppedNorman was first introduced to ACD/Labs software while working for Boehringer Ingelheim (BI) in Laval, Quebec (Canada) where he began his career working on the synthetic chemistry team. Once his comfort with computers was discovered, however, he was asked to work in the NMR Spectroscopy Lab (circa 1991, 3 years before ACD/Labs even existed). “At the time we had Bruker instruments with Bruker Unix based operating systems using the XWinNMR software”, Norman remembers. “I was later put in charge of building and maintaining a Linux cluster and began to spend 30-40% of my time on IT and the remaining time on structure verification and elucidation for samples submitted by my synthetic chemist colleagues.”

BI chemists had direct access to two 400 MHz instruments, while the 600 MHz and 700 MHz instruments were reserved for use by the specialized spectroscopists. Norman trained with Steven LaPlante—the resident NMR expert. Norman recalls, “Bruker’s TopSpin software was pretty expensive at the time, so we provided the chemists with WinNMR PC software to process their data offline from the spectrometers. Several years later, WinNMR was no longer supported and we started to look around for something new and better.” This was Norman’s first introduction to ACD/Labs. “We purchased 50 seats of ACD/Labs software for 1D NMR processing [at the time on the ACD/SpecManager platform]. The whole chemistry team used the software for basic processing, using the results to monitor ongoing syntheses.” Norman worked on the more difficult samples where structures were unknown, or with dilute samples that needed longer acquisition times for which he would run an expanded number of experiments (HMQC, HMBC, COSY, ROESY) to aid structure identification. 

“When ACD/Spectrus Processor was introduced in 2009 we upgraded our 50 licenses to the Spectrus Platform”, says Norman, “and I supervised the chemists training and use of the software.  I was still also working on small structure identification—analyzing compounds chemists were producing and providing structure confirmation. In addition, I was analyzing compound libraries (of 70-100 compounds) and that’s when I was introduced to NMR Expert [now known as Automated Structure Verification—this software solution automatically sweeps data from the instrument, followed by automated processing, and structure/spectrum verification based on provided expected structures].”

Norman has also been an influencer in the development of automated structure verification software (ASV). At first, only proton NMR spectra and expected primary structures were available for ASV. Using the first versions of the software, Norman found that approximately 60% of structure verification was reliably handled by the software. This was impressive at the time given that the primary structures of hundreds of compounds required elucidation. Nonetheless, it also required Norman’s manual analyses of too many compounds to avoid false positive and false negative conclusions. Norman and ACD/Labs were not satisfied with the current state-of-the-art. Together, they further improved the accuracy of ASV by incorporating chemical shift information of related compounds from internal experimental databases. “Ten representative samples or structures for a previous project were usually enough to make 50% improvements in ASV accuracy. A training database was important for good results.”

Further improvements became possible once ACD/Labs included the option of using HMQC NMR data in the ASV routine. This led to a serious reduction of inaccurate elucidations per library from 40 to 20-25, and as a result, it also significantly reduced Norman’s workload. When asked why he earlier persevered with ASV software even when it was only passing 60% of compounds, Norman responded, “It was advantageous for me because it still meant I didn’t have to analyze all of the compounds. The software was new at the time for me, and for ACD/Labs. I worked very closely with Arvin (an application scientist with a strong NMR background) and acted like a beta tester. I was good at finding bugs in the software, because I was using it in ways that ACD/Labs developers and application scientists would not even think of using it. It was a pleasure because I knew that my findings helped improve the software.”

Naubry-blog-FB

Norman took early retirement when the BI site closed in 2013 and has been working for Steven LaPlante, who  took a position at the Institut National de la Recherche Scientifique, ever since. Norman helps to train the research students on ACD/Labs software, and still assists in structure determination. “It’s really nice how with the internet and NMR Workbook Suite I can do everything off-line from the instrument. I work from home for personal reasons and can still keep my brain challenged.”

Steven recently purchased ACD/Structure Elucidator and Norman has been training with the Application Scientist team. “We‘re interested in using the functionality for 3D structure determination,” Norman explains. The recent version of Structure Elucidator (SE) enables interpretation of ROESY experiments while taking the constraints of NOE distances into account for efficient determination of 3D structure. “Without this software our workflow to determine the same information includes manually pulling values from the NOE table and inputting them into other software”, says Norman. “Now I’m hoping to be able to do that in SE so I can do everything in one software. I plan to again join the team as a beta-tester to make sure a good working strategy is found.”

Over the years that Norman has been using ACD/Labs software, he has transitioned from the SpecManager to Spectrus Platform and used a variety of different products. What advice does he have for people starting to use ACD/Labs’ NMR software?  “I’ve been using the software for so long that I find it fairly easy to transition, it’s very logical to me. My advice for someone picking up the software for the first time is to make sure the basic processing is correct—peak picking and integration. You have to respect the basics…those are important. If you don’t get the foundation right, the software will not give you the correct information. When students say the software doesn’t work, I ask them to show me what they did, the problems usually lie in improper peak picking and integration.”

Throughout his career Norman has worked with Bruker instruments so why not use the Bruker software? “I prefer to use ACD/Labs software because it’s what I’m used to now. Also, ACD/Labs personnel are quick to help if I find a bug. If I run into a problem with the software, it’s always good to be able to talk to someone with more experience or who understands it better. With ACD/Labs when there is a problem there is always someone there to help you. The software is very powerful which also means that sometimes it may take a little time to learn how to best use a particular feature. In all my time working with the software I’ve never encountered a time when a question went unanswered. Even if you didn’t have an immediate answer, I always get a constructive response. I have software I’m very happy with so why would I look elsewhere?”                                                                                                                                                                                              

Comments

Verify your Comment

Previewing your Comment

This is only a preview. Your comment has not yet been posted.

Working...
Your comment could not be posted. Error type:
Your comment has been posted. Post another comment

The letters and numbers you entered did not match the image. Please try again.

As a final step before posting your comment, enter the letters and numbers you see in the image below. This prevents automated programs from posting comments.

Having trouble reading this image? View an alternate.

Working...

Post a comment

Your Information

(Name is required. Email address will not be displayed with the comment.)