Microorganisms are ubiquitous on earth and play a prominent role in the material cycle, climate change, and human health (Lynch and Pedersen, 2016; Crowther et al., 2019; Jansson and Hofmockel, 2020; Keohane et al., 2020). In the last 2 decades, the development of high-throughput velocimetry allowed us to observe the structure of microbial communities, in which the Earth Microbiome Project (EMP) and the Human Microbiome Project (HMP) have achieved fruitful results (Turnbaugh et al., 2007; Human Microbiome Project, 2012; Gilbert et al., 2014; Thompson et al., 2017). In the last few years, a lot of studies on the interaction between various community microorganisms and their hosts have emerged (Lundberg et al., 2012; Ren et al., 2021). Numerous studies have discovered that the loss of gut microbiota homeostasis exerts a significantly negative impact on Alzheimer’s disease, obesity, cancer, and depression. (Dao et al., 2016; Zheng et al., 2016; Flemer et al., 2017; Kim et al., 2020; Morais et al., 2021). In addition to the metagenomics technique, the multi-omics technique has been used to explore specific interactions between microorganisms and hosts in a greater depth (Lloyd-Price et al., 2019). The Integrative Human Microbiome Project (iHMP) has conducted a series of integration studies to explore human microbial–host interactions from the multi-omics data in three physiological or pathological states of human prematurity, inflammatory bowel disease, and prediabetes (Integrative, 2019).

In multi-omics studies, there are two major challenges for researchers. One is that data integration and analysis consume considerable time and effort of researchers because of the complex usage of combining tools (Lin et al., 2020), and the other is that the existing drawing tools cannot satisfy the demands for describing high-dimensional data, which poses stress for the researchers to represent the analysis results of microbiome data (Sinha et al., 2015; Ramirez et al., 2018). Visualizing the results of data obtained from multi-omics studies is a huge burden for the researchers (Lin et al., 2020).

Currently, various tools have emerged and were used to analyze multi-omics data and visualize omics results. The main steps of the metagenome analysis are clustering or denoising the raw data to obtain abundance tables. In the analysis phase, QIIME2 (Hall and Beiko, 2018; Knight et al., 2018; Rai et al., 2021) performs better than other tools (Straub et al., 2020). However, users have to use dedicated tools to view graph files generated from QIIME2, which adds burden for the users to observe the results (Min et al., 2021).

For big data obtained from the multi-omics studies, visualizing the analysis results requires easier tools (Chen et al., 2020). Traditional graphing tools exhibited excellent graphing capabilities and convenience, especially client tools that are user-friendly but have problems such as low system compatibility and high costs. For example, OriginPro and GraphPad Prism v9 are two drawing tools for the general public, with rich styles of charts but of high economic costs. Command line-based tools, such as the Matplotlib package in Python and the ggplot2 package in R, provide novel graphing methods and graph styles with excellent plotting capabilities and are usually used for the secondary development (Skidmore et al., 2016; Wagih, 2017; Tareen and Kinney, 2020; Liu et al., 2021; Xu et al., 2021). A lot of command line-based tools which are cross-systems and open sources have been developed by bioinformaticians to exhibit the results of the multi-omics data (Ito and Murphy, 2013). However, those command line-based tools require users to spend considerable time learning a programming language, which decreases the efficiency for scientific researchers in non-computer fields, especially doctors and experimenters.

For these issues, we developed MicrobioSee, a web-based toolkit for multi-omics visualization, which is a cross-platform, user-friendly, time-saving, and an open source (Figure 1). The idea of this tool originated from the fact that conventional charts and tools cannot satisfy the demands of visualization for researchers without programming experience in their multi-omics studies. MicrobioSee is an efficient toolkit for visualization that eliminates high costs for users without programming experience.

MicrobioSee was developed for microbiome multi-omics data, such as the metagenome, proteome, and transcriptome (Figure 2). The whole website was divided into front-end and back-end. In the front-end, the VUE technology is used to render the interface. In the back-end, the R program was used for responding to users’ interactions and built-in drawing commands. In addition, most of the tools in MicrobioSee were built with shiny services to achieve real-time interactive plotting. A series of graphics could be plotted interactively by MicrobioSee, including the rose plot, heat map, box plot, upset plot, Venn diagram, and so on, which could be applied to multi-omics studies (Figure 3). The options for height, width, and resolution in each drawing module were provided for plotting.

To display the utility of MicrobioSee, three case studies were chosen and visualized by MicrobioSee. For brevity, we cannot explore all the functions of MicrobioSee but focus on the visualization of the most common scientific questions. Relevant data in case studies can be made available in the supplementary files.

In this work, we developed MicrobioSee, a web-based toolkit for the multi-omics studies, which contains plenty of computer technologies. The burden of plotting the result of the multi-omics studies would be eased with MicrobioSee. The pirate plot and raincloud plot are utilized as complements to the existing graph styles for visualization of alpha diversity results. Benefiting from the abundant R package resources, 17 plotting modules were developed for MicrobioSee. MicrobioSee also offers a few innovations in aesthetics. For example, elegant parabolic curves were added to the stacked column charts which would make them more aesthetically pleasing in the visualization of species composition.

The functions and advantages of the eight tools were summarized (Table 1). The vegan package (Oksanen et al., 2013) and phyloseq package (McMurdie and Holmes, 2013) are recognized by many researchers, but it is not friendly for researchers because of lacking interfaces. In the studies of the amplicon, QIIME2 (Hall and Beiko, 2018) and USEARCH (Edgar, 2010) are popular for high-speed analysis, but it is rarely used in the visualization of results on account of the insufficient number of their graphic styles. TBtools (Chen et al., 2020) has been popular with Windows users since it was developed, but it is disappointing for users of other platforms. Web-based tools could be used directly through a modern browser without platform limits. Animalcules (Zhao et al., 2021) provides an interface by Shiny technology, but it is a localized service with complex dependencies and could be installed with errors. Metaviz (Wagner et al., 2018) provides a web application for interactive visualization of the microbiome, but the methods of metaviz are not comprehensive or specific for 16S rRNA, metagenomic, or transcriptomic data. MicrobiomeAnalyst (Dhariwal et al., 2017; Chong et al., 2020), an excellent web toolkit in the field of downstream microbiome analysis, provides analysis and visualization, but few visual graphic styles are available. MicrobioSee was developed for interactive visualization of the microbiome, and microbiologists could use it for free and install it for free after short learning. The results in various omics studies can be visualized by MicrobioSee besides 16S rRNA and shotgun sequencing microbiome data. Due to its flexibility, it can also be applied in other fields. As long as uploading is in the same format as the sample data, the images of the corresponding graphic style would be obtained.

For a more convenient operation, the tool would be continuously updated. Compared with client tools, it is unnecessary to be reinstalled when a new version is released, which is user-friendly. Inconveniently, web tools are highly dependent on the web environment (Chen et al., 2020) and so is MicrobioSee. When the number of users increases to a certain extent, the servers and bandwidth for MicrobioSee need to be expanded. The servers of MicrobioSee may suffer from attacking for the global open access, and the firewalls need to be constantly upgraded.

In total, 17 plotting modules have been built for MicrobioSee so far, such as the violin plot, box plot, rose plot, heat map, box plot, upset plot, and Venn diagram. Although most functions are not unique to MicrobioSee, they were combined, optimized, and interfaced for researchers with limited coding experience. MicrobioSee simplifies the methods for users without programming experience.