News

Shimadzu Corporation has released CellTune software for supporting optimization of cell culture conditions that uses artificial intelligence (AI) to assist with selecting optimal cell culture conditions. This product suggests optimal conditions, such as the optimal culture medium composition or culturing environment, based on data from culture supernatant analyzed using a liquid chromatograph mass spectrometer (LC-MS) system and LC/MS/MS method package for cell culture profiling (Ver. 3).

Large-scale production of antibody drugs or cell-based drugs using cell culturing technology can involve significant risk, due to the time-consuming culturing processes and expensive culture media. Therefore, it is important to specify culture conditions (parameters) that can produce high-quality cells at a low cost. The main parameters involved in cell culturing typically include the concentrations of amino acids, vitamins, and other components in the culture medium and the temperature and pH level (hydrogen ion concentration indicator) of the culturing environment. On the other hand, culture media can contain about a hundred kinds of components, with different culture medium compositions and concentrations for each cell. Conventionally, such parameters were determined by experts with extensive experience, but that resulted in remaining efficiency problems. Therefore, since February 2022 Shimadzu has been working with an AI startup company Epistra Inc. (based in Minato-ku, Tokyo) to develop solutions for optimizing culture conditions. That collaboration resulted in the development of CellTune, which includes a feature extraction module for identifying parameters that have a major influence on culturing results based on data from analyzing 144 components in culture supernatant using a Shimadzu LC-MS system and an AI automatic optimization module for suggesting optimal culture conditions.

Beginning today, Epistra will use CellTune to offer a culture condition optimization consulting service for pharmaceutical companies, contract drug manufacturers, food manufacturers with microbial fermentation technology, and others. Shimadzu Diagnostics Corporation (SDC), a Shimadzu Group company, manufactures reagents, culture media, and other products related to cell culturing. In 2024, Shimadzu Corporation and SDC will launch a new business that generates recurring revenues from supplying culture media suggested by CellTune software. In the future, Shimadzu Corporation will also develop various technologies and products for assisting cell culturing based on using CellTune in addition to parameters based on measurement results from instruments other than cell imaging or LC-MS instruments.

This product is sold only in Japan and overseas sales are under preparation.

Features

1. Analysis Results can be Obtained without Statistical Analysis Knowledge

Culture media components that have a large effect on the culturing score can be identified by statistical analysis of data from LC-MS analysis of culture supernatants. Each processing program necessary for analysis is supplied while being linked via the software (analysis recipe), so that analysis results, such as identification of important components, can be obtained by simply specifying the analysis recipe and analysis parameter settings.

2. Suggests Optimized Conditions that are Difficult to Visualize

If multiple parameters are involved in determining the culture medium, then the optimal concentration is normally determined for one type of component before determining the optimal concentration for the next component. However, even if each component is analyzed using optimal parameter settings, it does not result in the best possible experimental conditions in some cases. In contrast, the AI automatic optimization module included in this product can evaluate multiple components at once to provide parameter settings optimized for the overall medium.

3. Searches for Experimental Conditions Efficiently

If the results (culturing score data) from repeating an experiment using experimental conditions suggested by CellTune are entered into CellTune, the software can generate a model for predicting culturing scores and then suggest the next experiment design based on that model. Operating efficiencies can be improved by conducting multiple experiments to increase the accuracy of predictions in accordance with the quantity of data entered.