Shiseido Develops Unique DNA Testing Method by Fusion of Dermatology Research and AI Technology

 Shiseido combines years of dermatological research findings and accumulated big data of 1472 Japanese women with AI technologies such as machine learning methods to clarify the relationship between DNA characteristics (SNP) and skin condition. did. In addition, we have succeeded in developing a new DNA test method that can evaluate the innate skin characteristics that differ from person to person with high accuracy by constructing a unique algorithm from the obtained results.

This test method contains 79 types of wide-ranging SNPs that include not only factors that have a direct effect on the components that make up the skin, but also factors that have an indirect effect on the skin from the body, such as blood vessels, hormones, and vitamin metabolism. By using an algorithm constructed by combining information from the above and a huge amount of actual skin measurement data, it has become possible to understand the customer's skin with high accuracy and holistic.

Aiming to realize new beauty care that is close to each customer by facing a huge amount of data on skin and DNA and deepening research on the relationship between the characteristics of DNA and skin conditions that differ from person to person.

It is known that skin condition is affected by innate genetic characteristics in addition to skin care including UV protection and lifestyle-related effects such as diet, exercise, sleep, and smoking. It is known that the genetic characteristics are influenced by the difference in SNP of DNA, and that each individual has various SNPs. If we can understand the characteristics of the born skin based on such SNP information in addition to the current skin condition, it will lead to the proposal of highly personalized beauty care that has never been seen before.

Therefore, the company combined many years of dermatology research with new technologies such as AI, and used big data of 1472 Japanese women to proceed with research on the relationship between SNP and skin condition.

First, according to the company's previous research, in addition to factors inside the skin that directly affect the skin condition such as collagen metabolism-related factors, it is indirectly involved in the skin such as blood vessel condition, nutritional component metabolism such as vitamins, and hormone metabolism. 79 types of SNPs were carefully selected from the suggested internal factors and set as analysis targets.

Subsequently, UV history, smoking history, and age information were added to the skin measurement data of abundant items such as wrinkles, stains, and barrier functions, and used for machine learning. We compared each item of skin condition by age group, analyzed the characteristic SNP of people who deviated from the average by 25%, and constructed a unique evaluation algorithm.

Then, using an original algorithm, we found a combination of 5 to 10 types of SNPs corresponding to each item of various characteristics of the skin that human beings are born with, such as "easiness of wrinkles / difficulty of wrinkles".

In a general skin DNA test, one or two types of SNPs are analyzed for each item of skin condition, but it was found that the accuracy is improved by analyzing more SNPs. Furthermore, by utilizing an algorithm constructed from a combination of various SNPs and a huge amount of actual skin measurement data, we succeeded in developing a new DNA test method that evaluates the characteristics of each individual skin with extremely high accuracy.