Analyzing metagenomes with HUManN, LEfSe, and GraPhlAn

Metagenome files can be submitted for further analysis to HUMAnN. HUMAnN takes gene abundances as inputs and produces gene and pathway summaries as outputs.

Input for HUMAnN

HUMAnN can be downloaded from the website. The input files need to be in the Tab-Separated Values (TSV: .tsv) format.

Converting QIIME tables to TSV format

QIIME tables can be converted to TSV using the script QiimeToMaaslin. To run the script, execute the following command

python qiimeToMaaslin.py metadata.metadata < inputfile.txt > outputfile.tsv

Providing the metadata is optional. For more information, please refer to the documentation.

Converting BIOM files to TSV format

BIOM files can be converted to TSV format using the tools provided by the biom format package (Please refer to the examples on the website for more information). Use the following command to convert biom files to TSV format

biom convert -i input_table.biom -o output_table.tsv -b

Please make sure to remove the first line in the output file (output_table.tsv):# Constructed from biom file, and save the file in the current format.

Running HUMAnN

The TSV input files can then be copied to the input folder in the HUMAnN repository (i.e. ../humann/input/). To execute HUMAnN, run the following command from the main repository path (i.e. ../humann/)

scons

This command will create an output directory in the main repository (i.e. ../humann/output), which will contain all the analysis results for each input file submitted.

Differential abundance analysis with LEfSe

Any of the output HUMAnN files named as: 04b-*-mpt-*.txt or 04b-*-mpm-*.txt can then be used as input for further analysis with LEfSe equivalently. Please follow the instructions below to make the input appropriate for LEfSe,

• Select a file from the HUMAnN output folder (named 04b-*-mpt-*.txt or 04b-*-mpm-*.txt)
• Open the file in Microsoft Excel or a text editor.
• Remove the first column.

• Remove every metadata row (anything including and above InverseSimpson) except the class (and optional subclass), and the top row: ID/NAME.

  • Please ensure only 1-2 metadata rows remain apart from the Name/ID row at the top.

• Save the modifications to the file, and use this version of the file as an input for LEfSe.

1. Load data with LEfSe by clicking on the Choose file button. Select the modified output file, and click on the Execute button.

2. Once the data has been uploaded (the file will appear on the right-hand-side panel), proceed with formatting the data for LEfSe by clicking the link Format Data for LEfSe in the panel on the left, and selecting the data from the drop-down menu.

3. Follow the instructions to select the correct fields in the drop-down menus, and then click on the Execute button.

5. Once the data is formatted, click on the LDA Effect Size link in the panel on the left. Select the formatted data from the Select Data drop-down menu and press Execute.

6. The output generated from the step above can be used as input for the following.

   • Plot the LEfSe results using the Plot LEfSe Results link in the panel on the left.
   • Plot a Cladogram using the Plot Cladogram link in the panel on the left.
   • Plot Differential Features using the Plot Differential Features link in the panel on the left.

Visualization with GraPhlAn

Either of the output files 04b-*-mpt-*-graphlan_tree.txt or 04b-*-mpm-*-graphlan_tree.txt can be used as input for further analysis with GraPhlAn equivalently.

1. Click the link Load input tree in the panel on the left, and select the output file from HUMAnN by clicking on the Choose file button. Press the Execute button to upload the file.

2. After the data has been uploaded, click on the Annotate tree link to add all the graphical features. Then, select the input file from the Input File drop-down menu. Specify the data fields according to the desired output, and press the Execute button when done. For example, the fields specified for a figure with leaf node names would be as follows:

   • Select the clades of interest from the list Select clade(s).
   • Enter * for the field Annotation Label.
   • Specify Clade leaf nodes from the drop-down menu Annotation Label Clade Selector.

3. Click on the Add rings to the tree link, and select the annotated data from the above step (instead of your raw input in Step 1) from the Input Tree drop-down menu. Upload the 04b-*-graphlan_rings.txt file (can be found under /humann/output/) through the Get Data link (located in the LOAD DATA MODULE in the panel on the left). Select the 04b-*-graphlan_rings.txt file from the Ring input File drop-down menu, and press Execute.

4. To plot the final tree, click on the Plot tree link in the panel on the left, and select the output from the step above. Press Execute. To visualize the results, click on the Eye symbol next to the output file generated in the panel on the right.