Integration of metagenomics and metaproteomics in the intestinal lavage fluids benefits construction of discriminative model and discovery of biomarkers for HBV liver diseases

Abstract

AbstractIntestinal lavage fluid (IVF) containing the mucosa‐associated microbiota instead of fecal samples was used to study the gut microbiota using different omics approaches. Focusing on the 63 IVF samples collected from healthy and hepatitis B virus‐liver disease (HBV‐LD), a question is prompted whether omics features could be extracted to distinguish these samples. The IVF‐related microbiota derived from the omics data was classified into two enterotype sets, whereas the genomics‐based enterotypes were poorly overlapped with the proteomics‐based one in either distribution of microbiota or of IVFs. There is lack of molecular features in these enterotypes to specifically recognize healthy or HBV‐LD. Running machine learning against the omics data sought the appropriate models to discriminate the healthy and HBV‐LD IVFs based on selected genes or proteins. Although a single omics dataset is basically workable in such discrimination, integration of the two datasets enhances discrimination efficiency. The protein features with higher frequencies in the models are further compared between healthy and HBV‐LD based on their abundance, bringing about three potential protein biomarkers. This study highlights that integration of metaomics data is beneficial for a molecular discriminator of healthy and HBV‐LD, and reveals the IVF samples are valuable for microbiome in a small cohort.