User data is not available to any other user* and confidentiality is maintained with the exception of jobs run on unsecured guest accounts (see Guest Accounts for more information).

* We try to provide a secure and private service for registered users but due to missed test cases or programming errors, some user data may be revealed unintentionally. We encourage users of this site to report such findings to the authors so we can remedy the problems as soon as possible.

Visibiome is best experienced with a large resolution screen. We recommend at least 1200px-wide screen for proper site rendering.

Users are not required to register to use the services. However, upon submitting an anonymous job, a guest account will be created for the user and will only be temporarily available. The password for any guest account is guest123. All guest accounts will be deleted within 24 hours of creation (along with any running or completed jobs). To prevent this from happening, simply update the details of the guest account. Please update the password since all guest accounts have identical passwords.

Visibiome takes BIOM file as an input. BIOM file can either be in Sparse or Dense format. It is designed to be a general-use format for biological sample by observation contingency tables. BIOM is a recognized standard for the Earth Microbiome Project. Users can use Qiime to create a BIOM formatted table after running closed-reference OTU picking on sequences against GreenGenes 13.5.

Sample size limitations allows up to 100 samples for Bray-Curtis and AESA/UniFrac searches while only up to 10 for GNAT/UniFrac in the interest of timely computation. To compare more user samples, please find a download link to the VirtualBox image to run a personal deployment of Visibiome and adjust the maximum number of samples as needed in the settings/ folder.

Example TSV text:

# Constructed from biom file
#OTU ID     S1
4371191     0.0
126871      3.0
533625      5.0
217471      4.0
834306      5.0
114076      6.0
253428      5.0
544561      8.0
813711      9.0
1827890     1.0
4294883     2.0

Example JSON text (output from biom-format.Table object using the as_json() method):

  "id": "None",
  "format": "Biological Observation Matrix 1.0.0",
  "format_url": "",
  "matrix_type": "sparse",
  "generated_by": "Qiime 1.8.3",
  "date": "2016-07-17T12:36:44.034878",
  "type": null,
  "matrix_element_type": "float",
  "shape": [11, 1],
  "data": [
    [1, 0, 3],
    [2, 0, 5],
    [3, 0, 5],
    // ...
  "rows": [
      "id": "253428",
      "metadata": null
      "id": "533625",
      "metadata": null
      "id": "1827890",
      "metadata": null
    // ...
  "columns": [
      "id": "S1",
      "metadata": null

Visibiome has several ways to put the user samples in an environmental context. Mainly, the contextualization is done through visualizations of a queried sample against the Visibiome database samples. These visualizations are generated using d3.js when the user loads the page. For the optimal experience, we recommend that users download a modern browser which is capable of viewing SVGs and on a computer with enough memory to support the large visualizations. Typically a recent version of Mozilla Firefox or Google Chrome is more than adequate. IE9 and below are unsupported and no guarantees can be made for the visualizations appearing properly in those browsers.

Available visualizations:

  1. Ranking with barcharts
  2. Dendrogram against representatives
  3. PCoA plots against representatives

Not all analysis types feature the same visualization options due to the limitations of the algorithm in order to trade off for a faster search speed. We apologize for the missing visualizations and we are actively adding new features to accomodate the shortcomings.

Ranking has 2 different ways to present the most similar samples in our database to the user sample based upon the distance measure selected by the user when making a query: clade barcharts and ranking cards.

Visibiome features 3 barcharts, for three different taxonomic ranks: phylum, family genus. Each barchart contains clades of the samples which are matched (including the user-submitted samples) according to each taxonomic rank. These clades are inferred based on the GreenGenes ID of the OTUs in the samples. The barcharts allow the users to identify the culprits of what is causing the samples to match and a higher granularity compared to other plots in Visibiome.

The ranking cards allow users to view the most similar samples to the uploaded samples. It contains the details for each sample such as study title and source, EnvO(s), Ecosystem, and Sample depth. These details help user to characterize uploaded samples against actual samples which can be found in popular microbiome databases and perform more specific downstream analysis.

Why do I see "No barcharts available for this sample" in my ranking page?

  1. We do not currently provide barcharts for Bray-Curtis queries.
  2. Our algorithms have strict bounds on distances and if matched samples are much farther than a threshold, we do not consider the matched samples as a relevant match and no barcharts are drawn for these user samples. When relevant, a maximum of 6 matches (bars) are shown. Poorly matched results are still placed in the ranking table.

A dendrogram is a tree diagram used to illustrate the arrangement of the clusters produced by hierarchical clustering. Each sample has colored bar(s) to represent the ecosystem(s) it belongs to, simplifying the environmental similarity to the user. Hovering the mouse over the leaf nodes will show more details about the sample and one or more EnvO label of the samplefor a more comprehensive visualization.

Principal coordinates analysis (PCoA) adds another dimension to contextualizing the samples. It is hard to visually see how the samples cluster together in high dimensional space. Using PCoA, we reduce the dimension of the original data down to 2 dimensions split into three plots of top 3 most "important" principal coordinates. The PCoA plots gives a rough idea how user samples cluster against the Visibiome database samples and as well as how all the matches cluster together.

If you wish to run an example task to be familiar with the user interface, you can view an example result of a pre-run job.

Download Visibiome

We offer a prepared copy of Visibiome (excluding a few databases to minimize storage size) for users to download and use. Visibiome is distributed as a VirtualBox image under Xubuntu 16.04. The download link can be found here and further information regarding how to download the missing files are attached in a README file in the distribution.

To install Visibiome,

  1. extract the .vdi file from the downloaded file,
  2. Once extracted, run the VM VirtualBox Manager on your computer,
  3. Create a new virtual machine and a window will appear,
  4. Give the new VM an appropriate name, and set the type to Linux and the version to Ubuntu (64-bit). Once completed click next,
  5. We recommend setting the memory to at least 2048 MB (2GB) although more is advised. Once completed click next,
  6. For the hard disk, select User an existing virtual hard disk file. Select the extracted .vdi file and click next,
  7. Click finish and you should be able to run the Xubuntu VM. Open a new web browser window and Visibiome should appear on the homepage.

We urge users to read the before-you-start.txt file on the desktop before continuing to use Visibiome as some files are still missing from a complete installation of Visibiome and needs to be downloaded. It also contains important information regarding passwords and access details. Please report any errors you may encounter to the authors of Visibiome or to the git repository. Contact information can be found in the contact page.