The data were acquired by UNIFI software (Waters) and processed by UNIFI and Progenesis QI software. Omics analysis methods, such as principal component analysis (PCA), partial least squares-discriminate analysis (PLS-DA), and orthogonal partial least squares-discriminate analysis (OPLS-DA), were employed to identify the biochemical patterns in tongue coating samples and suggest variables that can be used as biomarkers for diabetes and pre-diabetes.

The primary outcome is the differentiation of tongue features extracted by ATDS. There are nine primary features for TCM clinical tongue diagnosis, namely, tongue shape, tongue color, fur thickness, fur color, saliva, tongue fissure, ecchymosis, teeth mark, and red dot. Features identified are further sub-divided according to the areas located, i.e., heart-lung area, left liver-gall, right liver-gall, spleen-stomach, and kidney.

The study researched tongue coating metabolic markers in patients with diabetes to determine any relationship between tongue diagnosis and metabolic processes. Chromatographic analysis and comparison between the diabetes group and normal controls. PCA and PLS-DA analysis of metabolic profiles of tongue coating samples from the chronic gastritis group and normal controls. To improve the accuracy of the PLS discriminated model, OPLS-DA was used to analysis the results by removing some redundant information, such as envir- onmental factors, gender, and diet. Tongue metabolic fingerprint differences between the chronic gastritis group and normal controls were compared using OPLS-DA.