Discover Oncology (Jan 2023)

Analysis of bacterial diversity and community structure in gastric juice of patients with advanced gastric cancer

  • Qiang Wei,
  • Qi Zhang,
  • Yinhang Wu,
  • Shuwen Han,
  • Lei Yin,
  • Jinyu Zhang,
  • Yuhai Gao,
  • Hong Shen,
  • Jing Zhuang,
  • Jian Chu,
  • Jiang Liu,
  • Yunhai Wei

DOI
https://doi.org/10.1007/s12672-023-00612-7
Journal volume & issue
Vol. 14, no. 1
pp. 1 – 16

Abstract

Read online

Abstract Background The occurrence and development of gastric cancer are related to microorganisms, which can be used as potential biomarkers of gastric cancer. Objective To screen the microbiological markers of gastric cancer from the microorganisms of gastric juice. Methods Gastric juice samples were collected from 61 healthy people and 78 patients with gastric cancer (48 cases of early gastric cancer and 30 cases of advanced gastric cancer). The bacterial 16 S rRNA V1-V4 region of gastric juice samples was sequenced. The Shannon index, Simpson index, Ace index and Chao index were used to analyze the diversity of gastric juice samples. The RDP classifier Bayesian algorithm was used to analyze the community structure of 97% OTU representative sequences with similar levels. Linear discriminant analysis and ST-test were used to analyze the differences. Six machine learning algorithms, including the logistic regression algorithm, random forest algorithm, neural network algorithm, support vector machine algorithm, Catboost algorithm and gradient lifting tree algorithm, were used to construct risk prediction models for gastric cancer and advanced gastric cancer. Results The microbiota diversity and the abundance of bacteria was different in the healthy group, early gastric cancer and advanced gastric cancer (P < 0.05). The top five abundant bacteria among the three groups were Streptococcus, Rhodococcus, Prevotella, Pseudomonas and Helicobacter. Bacterial flora such as Streptococcus, Rhodococcus and Ochrobactrum were significantly different between the healthy group and the gastric cancer group. The accuracy of the random forest prediction model is the highest (82.73% correct). The bacteria with the highest predictive value included Streptococcus, Lactobacillus and Ochrobactrum. The abundance of bacteria such as Fusobacterium, Capnocytophaga, Atopobium, Corynebacterium was high in the advanced gastric cancer group. Conclusion Gastric juice bacteria can be used as potential biomarkers to predict the occurrence and development of gastric cancer.

Keywords