Author - mestrelab

A successful collaboration with Novartis on automated analysis of ligand-observed 1H-19F NMR binding data for fragment based lead generation

Dr. Andreas Lingel
Novartis Researcher


Company: Novartis
Market: Pharmaceutical
Problem: Several groups in Novartis use NMR binding assays to find and validate hits during the early stages of drug discovery. They use ligand-observed NMR experiments such as STD, T1ρ, and CPMG to identify library compounds that interact with a target protein. Until a few years ago, the analysis of a high number of complex spectra was performed largely manually and therefore represented a limiting step in hit generation campaigns. Typically, manually stacking and visual analysis hundreds or even thousands of spectra can take multiple days and the results can be subjective and error-prone.


We have collaborated with two groups in Novartis since 2011 to develop a new plugin in Mnova for processing and analyzing ligand-observed proton and fluorine NMR binding data in a fully automated fashion. Scientists at Novartis evaluated the performance of this plugin by comparing automated and manual analysis results on 19F and 1H-detected data sets, and found that the program delivers robust, high-confidence hit lists in a fraction of the time needed for manual analysis and greatly facilitates visual inspection of the associated NMR spectra. These features enable considerably higher throughput, the assessment of larger libraries, and shorter turn-around times.


Mnova Screen and Mnova NMR


The scientists at Novartis found that using the Mnova Screen program can be faster manually analysis. For a typical screening data set, the computational runtime needed for the automatic analysis ranges from a few minutes up to 3 hours, while manual analysis can take half a day to multiple days. The automatic analysis provides robust and reliable results that overlapped with the manual analyses with an overall of higher than 90% agreement. Moreover, for the analysts, the organization and display of screening spectra stacked with all the reference spectra are of invaluable help for validating results and making decisions in cases where compound signals are close to each other.


Journal of Medicinal Chemistry: Fast and Efficient Fragment-Based Lead Generation by Fully Automated Processing and Analysis of Ligand-Observed NMR Binding Data. Chen Peng,*,†,∥ Alexandra Frommlet,‡,∥ Manuel Perez,† Carlos Cobas,† Anke Blechschmidt,‡,§ Santiago Dominguez,† and Andreas Lingel*,‡

†Mestrelab Research S.L., Feliciano Barrera 9B − Baixo, 15706 Santiago de Compostela, Spain
‡Novartis Institutes for BioMedical Research, 5300 Chiron Way, Emeryville, California 94608, United States

Carnegie Mellon University and Mspin


Dr. Roberto Gil
Research Professor and Director
NMR Facility


  • Institution: Carnegie Mellon University
  • Market: Academia – Department of Chemistry
  • Problem: Three tetranortriterpenoids were isolated as hemiacetals (the C-23 epimeric xylorumphiins M− O (46 )), which could not be purified due to the equilibration of the hemiacetal epimers in solution, hence giving complex spectra. In order to enable purification, compounds were acetylated and the R and S epimers could be separated. However, for these C-23 epimers, use of standard NOESY and ROESY techniques did not permit their differentiation, due to long distances between H-23 and protons from the main skeleton.


The relative configuration of the new triterpenoids was determined by Dr. Gil and coworkers using the MSpin RDC module. 1DCH RDCs were collected for the C-23 epimers 5a and 5b on a compressed PMMA gel by using HSQC-F1 coupled experiments.  The MSpin program was fed with the experimental RDCs and the computed structures for the possible C23 diastereoisomers and the relative configuration of both epimers of the molecule was established on the basis of MSpin RDC fittings. The RDC technique proved to be essential for the determination of configuration in the remote C23 side-chain center. Once the configuration at C-23 of each epimer (5a and 5b) was unambiguously determined by RDCs, the configuration of the related compounds (4a, 4b, 4c, 4d and 6a) was determined by direct comparison of 1H and 13C NMR chemical shifts.


Waratchareeyakul, W.; Hellemann, E.; Gil, R. R.; Chantrapromma, K.; Langat, M. K.; Mulholland, D. A. Application of Residual Dipolar Couplings and Selective Quantitative NOE to Establish the Structures of Tetranortriterpenoids From Xylocarpus rumphii. J. Nat. Prod. 2017, 80, 391–402.

Automating data delivery and databasing – Case Study

Top Pharma Company
Assistant Director


Institution: A top Pharma Company
Market: Chemistry & Pharma
Number of Employes: 200 discovery and development chemists
Uses for tool: Circa 100,000 sample per annum
Market: Third party ELNs (Biovia, IDBS)
Type of installation: In-house developed, long standing tool for data delivery to chemists (upkeep – maintenance/bug fixing a challenge)


  • Short term: To greatly simplify access of all chemists to their analytical data, review of the data and re-storage.
  • Mid term: To enhance the use of analytical data and avoid work repetition through lack of awareness or visibility of previous work.


The solution implemented has several elements:

  • MnovaMnova Database
  • Databasing automation
  • Workflow driven data access and review


The workflow is very simple and unified / standardised. It is therefore easy to teach new chemists and has little room for error / problems.

Improvements on previous system:

  • Data are databased in real time, automatically (no errors, nothing is missed)
  • Data are available to chemists as soon as acquired
  • Data are processed and templates applied in Real Time
  • Sample Number, Experiment ID, User ID, date and all other relevant metadata are harvested in Real Time and made available for searching
  • PDFs are generated and archived in Real Time
  • Search the system from a greatly simplified interface, within Mnova
  • Data are retrieved by Sample Number, Experiment ID and other clear identifiers
  • Data can be searched by date, user, project, etc., if ‘lost’
  • Structure, peak and multiplet search are also available
  • Chemist can reprocess and, after reprocessing, the following happens with one single click:
    • Resetting of reports – Apply templates
    • Resave to DB with new version number, tagged as human reviewed
    • PDF generated on the fly and archived

Improvements on previous system:

  • Time saved
  • No browsing around file systems to find data, just enter identifier
  • Easily find ‘lost’ data (if you don’t know where it is, the DB will find it)
  • Retrieve already analysed data with one click
  • Reprocess, reanalyse and reset layout of data with one click
  • Resave to DB, version and archive PDF with one click
  • No need to specify save locations, the system knows
  • Better information
  • All data for a given identifier (e.g., Sample Number) are presented to chemist, so he can see analytical data or structure, other chemists analysis, etc., which he may not have been aware of

The chemist workflow in detail



  • 1 min saved per chemist per dataset opened (not browsing complex file directories, just enter unique ID)
  • 1 min saved per chemist at report layout (no need to layout report or report any parameters or information, the system knows what it must report)
  • 1 min saved per chemist at saving and databasing (laying out report, saving processed and laid out file, saving to DB, generating PDF, saving PDF to specific file location are all done with one single click – one click for all operations)
  • 5 datasets per chemist per day (average number of samples generated by chemist in a day)
  • Fully loaded cost per chemist: €250,000 per annum (€1,000 per working day)


Economic value:


  • Followed a workflow analysis – design specification process / several iterations, 1 month
  • Customization and implementation development – 1 month
  • Alpha testing by acceptance team, including bug fixing and iterations – 1 month
  • Deployment: this requires a packaging process and for this size installation and given the customer resource dedicated to it takes about 1 month


  • Mnova Suite (x 200, 1 per chemist)
  • Processing Real Time Solution (x12, 1 per instrument)
  • Mnova Spectral DB (Server x2 for production and replicate, client x 200, 1 per chemist)
  • Smart Analysis plugin (x200, one per chemist)

Department of Chemistry – University of Wisconsin-Madison


Dr. Heike Hofstetter
Asst Dir NMR Lab – CIC


Institution: University of Wisconsin-Madison (US), Department of Chemistry
Number of Employes/Students: More than 300 users
Market: Academia – Chemistry
Problem to solve: NMR data processing software that can be used in an educational environment for undergraduate and graduate education, and also be a primary tool for high-end research.


Why Mnova NMR?
Mnova has proven to combine ease-of-use with very powerful features.  Undergraduates quickly learn how to do simple 1D-processing, and migrate smoothly into performing structure assignments and data work-up for coursework and then into research publishing.


” Undergrads generally like the program and find it easy to use – which might mostly be related to the fact that when you drag in FIDs, Mnova automatically converts them in the standard setting. It seem intuitive enough that they feel comfortable after only a short introduction. We get more comments about installation than actual use.

Most of our grad students have either used Mnova themselves or are using it in classes they teach. They think that they know the program and are then amazed about all the features they did NOT know about….  It speaks to these tools’ ease-of-use and practical utility that the students then constantly use these features once they do about them… e.g. the assignment tools. These students also like that they can change the setting to match requirements for journals, dissertations, etc.

In general the students also really appreciate that once the program is installed they will be able to use it not just for one semester but throughout the rest of their studies, including research, and for different classes. This saves them time of re-learning the basic usage when working on new assignments (after organic they should have a fairly solid base).”

Dr. Heike Hofstetter

Automated analysis and reporting LC/MS and NMR together for compound registration – A case study


A pharma company implements quality control of their compounds using both NMR and LC/MS before registration.

The analysis are generated in each individual format but they must consolidate these analysis.


Based on their requirements we have developed a Mnova script (R_NMR_MS.qs) which does all of this automatically.

The user opens an LC-MS and/or 1D 1H NMR spectrum, plus a structure in Mnova. Then he runs the script and Mnova will do the following analysis and reporting:

  • For NMR, it reprocesses the spectrum to make sure the apodization is right and baseline is corrected. Then it does peak picking and multiplet analysis. It displays the parameters table, the multiplets analysis results in a journal format, and shows two expansion areas of the spectrum (11 to 9ppm, and 5.5 to -0.5 ppm).
  • For LC-MS or GC-MS, it does the molecule match for both the positive and negative-polarized MS, and finds the retention time where the molecule ion and isotope peaks are bests matched. Using that retention time, it locates the UV peak in the PDA and reports its area as the UV purity.

The following screenshots shows the report with both NMR and MS in the same document:

One MS Solution

Solution for a general pharma company.


“We are fully satisfied with Mbook as we are using it from several months. We indeed recommend it: Mbook is a real open environment and Mestrelab itself is keen to help customer to develop solution that really suites their needs. So we are really looking forward to new developments.“

Michele Lombardo

Assistant at Industrialisation Process – Scientific Division


Company: Helsing provides licensing of pharmaceuticals, medical devices and nutritional supplement products, with focus on cancer supportive care to benefit both patients with cancer and caregivers.
Problem: The main objective was to transfer our routine which was based on a heavy tedious paper-based system, to a paperless solution, with all the advantages that a digital system could have.


Why mbook

We choose Mbook because of several reasons. First of all we had a very positive feedback being able to immediately have access and “play” with a free trial version of Mbook, which was something not so common from other vendors and it is indeed a plus. Our interactions with the standard Mbook version was really good, since it is very user friendly and well suited for an R&D lab. Besides that, the customization we have requested were in general very well welcomed and the overall quality of the product versus the price (product price plus customization) was really appealing.

Finally we have a very positive interaction with Mestrelab’s representative and development team. appealing.

About regarding customization

We were very impressed by the availability of Mestrelab’s representative taking into account all our requests. The key point here was the capability to translate our paper-based workflow into suitable tools in Mbook. Furthermore we positively appreciated the possibility to get developers involved at the very beginning of this process.

Using Mbook

We had a first meeting in May 2015 when we then had access to the free trial. By mid June we agreed a number of licenses and customization, and we were able to follow it step by step having access to beta version of our customization. Starting from October we installed Mbook locally on our server and we started a three months evaluation period with a small group of users. Since January 2016 we have extended access to Mbook to all 15 users.

Overall impression

We have quickly appreciated the customization we have requested. Implementation had an immediate positive impact on our workflow keeping the original user friendly interface of Mbook.


Our workflow now is indeed more rapid and effective, we have now a full and complete control of our compound database. Since Mbook is web based it is really easy to access to it from everywhere, either when we are working in the lab or being busy in a meeting. Talking in terms of productivity we decided to go further and equipped each fumehood with a tablet, hence Mbook is now used on a real time base.

We have so far estimated a strong reduction in time spent to get a new synthetic step started, a reduction in the amount of paper used and time spent to prepared laboratory reports, print them and classify them for further use.