The first result of this partnership is the newly developed BiologicsHOS software, which provides an efficient workflow for evaluating the higher order structure fingerprint of proteins in biopharmaceutical research. BiologicsHOS uses advances in data acquisition and analysis to characterize intact biopharmaceuticals at natural abundance, using both 1D and 2D NMR methods. Because of its intrinisically high information content, NMR is proving to be a uniquely valuable tool in the validation of HOS, reducing the number of techniques needed to characterize biologics and biosimilars.
A second example is the integration of Bruker’s Fragment-Based Screening (FBS) solution with Mestrelab’s MScreen software, which provides drug discovery groups with an integrated experience from data acquisition and analysis to the identification of hits in FBS-by-NMR campaigns.
Intelligent NMR Spectrometer Initiative
Bruker is continuing to advance the Intelligent NMR Spectrometer framework at ENC 2019 for ease of use, lower training requirements and higher productivity.
The new SmartDriveNMR acquisition tool utilizes a fast scout experiment along with all other user inputs, and determines if follow-up experiments are beneficial, in which case optimized further experiments are carried out in automation within the time limits set by the user. The NMR experiment portfolio includes heteronuclear 2D experiments including HSQCs and HMBCs, 1D 13C and a variety of different types of solvent suppression schemes for 1D 1H experiments. Non-Uniform Sampling (NUS) data acquisition and Signal-to-Noise optimization ensure data quality, and an automatic structure verification (ASV) is completed as an integrated part of SmartDriveNMR.
Automatic calibration for the Intelligent NMR Spectrometer is now offered by the new AutoCalibrate™ capabilitiy, which determines the optimal settings for key NMR parameters, logs results and monitors changes. AutoCalibrate is key to maintaining well-tuned NMR spectrometers, tracking changes and monitoring the long-term performance and health of the entire system.
Deep Learning Applications in NMR Spectroscopy is part of our work-in-progress project around Deep Neural Networks (DNNs). It now enables interested beta customers to use supervised learning, e.g. for the automatic detection of signal regions in 1D 1H NMR spectra. The deep neural network was trained on two million spectra that were simulated with artificial noise and other artifacts from publically available structures.