MicroEXPERT: Microbiome profiling platform with cross-study metagenome-wide association analysis functionality

Abstract

The framework of the MicroEXPERT platform. Our Platform was composed of five modules. Data management module: Users upload raw data and metadata to the system using a guided workflow. Data processing module: Uploaded data is processed to generate taxonomical distribution and functional composition results. Metagenome-wide association studies module (MWAS): Various methods, including biomarker analysis, PCA, co-occurrence networks, and sample classification, are employed using metadata. Data search module: Users can query nucleotide sequences to retrieve information in the MicroEXPERT database. Data visualization module: Visualization tools are used to illustrate the metagenome analysis results.

Figure 1

Screenshots and examples of user cases in MicroEXPERT (https://microexpert.aimicrobiome.cn/). (A) Homepage of the MicroEXPERT. Interfaces to other modules are supplied, including the introduction and the statistical information of MicroEXPERT. (B) Projects classified by biome information. The home page exhibits the biomes that hold the most projects. The latest submitted projects are also shown. (C) A screenshot of the biome information. All of the samples are classified into 235 biomes, each of which links to the project's details. (D) The example of sample information page. For each sample, the quality control result of the upload sample was illustrated. Low‐quality sequences were filtered out by steps. Moreover, the functional results were divided into three categories: biological process, Cellular Composition, and Molecular functions.

Figure 2

A screenshot of the MWAS workflow (https://microexpert.aimicrobiome.cn/mwas). (A) Sample selection. Users can choose all the samples processed by the analysis workflow in the MicroEXPERT database. Cross‐study selection is applicable. (B) Custom analysis selection. In this step, the users can check the selected samples and choose the analysis tools (MWAS analysis and comparison methods). At least 10 samples are recommended for MWAS analysis. (C) Metadata categorization. For the selected samples, their metadata in our database are uploaded. The users can categorize these samples into at most four groups labeled with the chosen color. Alternatively, users can upload the group information by importing the table from your local computer. (D) Sample comparison. The sample is compared using different statistical tools such as a principal component analysis (PCA) analysis, a heatmap, and stacked columns. (E) MWAS analysis. Based on the categorization information and taxonomical or functional composition, the association between the metagenome and metadata is analyzed using variety of statistical tools such as heatmap, mental test, PCA, and network.

Figure 3

Investigation the gene‐species‐sample relationship with one sequence (https://microexpert.aimicrobiome.cn/search). (A) Gene annotation workflow. For the query nucleotide sequences, the MicroEXPERT platform searches the genes in the NCBI nucleotide database using the tool BLAST+ (version 2.5.1). The detailed matched information based on the NCBI nucleotide database are also illustrated. (B) Species identification workflow. To identify the possible source species of the query sequence, Kraken2 (version 2.0.7) and MetaPhlAn4 (version 4.0.2) are used. The pedigree‐based and phylogenetic relationship of the matched species is deduced using the NCBI taxonomy database. (C) Sample search workflow. To explore the distribution of samples and the biome of the query species in the database, all analysis results and sample information are stored in a separate table to return the abundance information of the query species in different environments and samples. The samples containing the queried species are illustrated according to their relative abundance in the related samples.