Update publication information. Although a variety of omics studies from different tissues have been used to analyze periodontitis pathology and progression, lacks accurate signature molecules for periodontitis. This study aims to identify the potential signature molecules suitable for distinguish of patients with periodontitis using TMT-proteomics and artificial neural networks analysis. Gingival tissues were collected from one site of 15 systematically healthy individuals, in which 1 individuals have stage IV periodontitis and 4 have stage III periodontitis (Severe Periodontitis, SP), 5 stage II periodontitis (Mild Periodontitis, MP) and 5 periodontally-healthy (H). we first time carried out the protein profiles of gingival tissue in periodontitis and healthy individuals utilizing the quantitative TMT-proteomics and discovered 9 signature proteins involved in periodontitis by integrated analysis of proteomics and transcriptomics using LASSO and ANN analysis. This study provides a novel insight for potential signature molecules in gingival tissue to predict periodontitis.