What’s New in Spectroscopy?

“Pimp my spec” – an upgrade for NMR spectroscopy. NMR spectroscopy is a vital method for describing the atomic structure of biomacromolecules in their native solution state – but this method suffers from low sensitivity. So, how did researchers from the University of Vienna, Austria, go about solving this problem? They used hyperpolarized water in a method dubbed “Dissolution Dynamic Nuclear Polarization” to achieve a 1,000-fold signal amplification in NMR measurements, measuring biomolecules at concentrations as low as 1 micromole/liter. In the press release, one of the co-authors, Dennis Kurzbach, compared NMR spectroscopy to an electric guitar. “If the amplifier is too weak, you will hear very little if you do not hit the strings strongly," he said. "Meaning that you need a lot of material to see an NMR signal. With the new hyperpolarization amplifier, you can now see something even at low concentration".

Drink up. Green coffee beans currently go through rigorous quality control measures with experienced analysts to determine whether the coffee is considered “special” or “traditional.” However, there have recently been suggestions that testing should have limited human input for maximal consistency.  As such, researchers have developed a method combining multispectral imaging with machine learning models to distinguish specialty coffee from traditional coffee. One of the four algorithms tested, dubbed “SVM,” showed exceptional performance, with an accuracy of 0.96. The authors concluded that the use of multispectral imaging combined with machine learning is an effective method for screening special and traditional green coffee beans, with rapid results that do not require pretreatment or destruction of samples.

A promising breast cancer diagnosis method? Current methods of breast cancer diagnosis are time-consuming, costly and often dependent on physician experience. In search of a more reliable diagnosis method, a group of  researchers have come up with a promising diagnosis method combining serum Raman spectroscopy and different classification algorithms. Analyzing the serum Raman spectra of 171 invasive ductal carcinoma patients and 100 healthy volunteers, the researchers built four classification models: support vector machine (SVM), decision tree (DT), linear discriminant analysis (LDA), and Neural Network Language Model (NNLM). The LDA, SVM, and NNLM algorithms achieved 100 percent accuracy. The results demonstrated high reliability and validity in diagnosing breast cancer.