News

Shimadzu Corporation has released eMSTAT Solution Ver. 2.0, statistical analysis software that supports the analysis of chromatograms^*1 and direct ionization mass spectrometer data. This product is the first statistical analysis software in the industry for chromatograms^*2 to feature machine learning capabilities. Machine learning allows the creation of discriminant models for grouping and enabling the identification of unknown components. The software uses analysis data obtained by a liquid chromatograph (LC), gas chromatograph (GC), and MALDI (matrix-assisted laser desorption/ionization) type mass spectrometer (MS). Since multivariate analysis^*3 can be easily performed, differences between target groups can be clarified and shown in comparisons with other products during product development in the food sector or in determining the quality of products during quality control assessments. Shimadzu plans to market this software in areas such as the food industry, where there is a growing demand for multivariate analysis for analyzing functional ingredients, evaluating food taste, and so on, to contribute to the advancement of food tech by improving the efficiency of data analysis.

In the food sector, there is a demand to easily identify those components that affect differences in flavor and texture and to repeatedly use data about defective products found during shipment inspections to analyze the various trace components and complex composition of samples. The software features two modes: a Statistical Analysis Mode for easy visualization of differences and a Discriminant Analysis Mode that identifies unknown samples with models created through machine learning. Shimadzu used this software in its original sake brewing project to clarify the distinctive characteristics of Shimadzu's original sake. This included identifying the yeast-derived isoamyl acetate and isoamyl alcohols that form the aroma and the presence of higher levels of organic acids such as malic acid and succinic acid that are more abundant than in other Japanese sake.

Shimadzu Corporation has identified healthcare, including the life science field, as one of the key areas for creating social value in its medium-term management plan and is focusing on solving social challenges in this field. We aim to continue contributing to the realization of a healthy and long-lived society by providing various analytical and measurement technologies in the life science field, including food tech.

Window of eMSTAT Solution Ver. 2.0

• *1 Chromatogram: Measurement data that is the signal waveform obtained by a chromatograph
  *2 As of June 2024. According to Shimadzu investigations
  *3 Multivariate analysis: A general term for a statistical method to examine the correlation between multiple types of data (variables) obtained from a sample

Features

1. Statistical Analysis by Intuitive Operations

In Statistical Analysis Mode, components that characterize taste and aroma can be searched for intuitively through window operations. A dynamic grouping function allows samples to be flexibly grouped to visualize differences in characteristics among samples and compare component quantities on various evaluation axes. None of these functions require complicated parameter settings and can be performed almost entirely with mouse clicks.

2. Unknown Samples can be discriminated Using Machine-Learning Models

In Discriminant Analysis Mode, a machine-learned discriminant model can be created using analysis data of known samples. By simply switching modes, analysis data obtained in Statistical Analysis Mode can be used as-is in Discriminant Analysis Mode. Loading data from unknown samples into the discriminant model helps determine the group each data belongs to, which is useful for determining the type and origin of food products and making quality control decisions. This is the industry's first statistical analysis software for chromatograms that includes a machine learning function.

3. Even More Convenient Using Related Software

Combining this software with other Shimadzu software programs simplifies analysis even further. For example, by using this software after measuring components in a sample with the LC-MS software LC/MS/MS method package for primary metabolites, users can easily compare the content of major components like amino acids and organic acids in food, search for characteristic components, and discriminate unknown samples.

For more details, visit
eMSTAT Solution Ver. 2.0