Web-Based FunGene Pipeline


The FunGene Pipeline is for processing functional gene sequences. It differs from the processing of rRNA gene sequences in several respects:

  • Chimeras are usually removed using uchime in de novo mode.
  • FrameBot is used to correct sequencing errors (insertions & deletions) and translate DNA sequences.
  • FrameBot can output an OTU table based on closest match to a reference database. (Analogous to a supervised approach.)
  • HMMER3 is used to align the translated sequences.

The aligned sequences can then be processed according to the unsupervised method, following the same steps, except for classification, as used for rRNA gene sequences.

Using the Web-Based FunGene Pipeline

RDP’s FunGene Pipeline page is available at http://fungene.cme.msu.edu/FunGenePipeline/. Before running the Pipeline, you must have separate sample fasta or fastq files which can be obtained using the Initial Processing Tool. Clicking on “Pipeline” in the upper left of the page takes you to a page for running the standard pipeline. Fill out the form with your email address, job name, and select the gene from the drop down menu. Upload your nucleotide sequence file(s) from the Pipeline Initial Process step. Sample sequence files may be zipped together for upload. Check the appropriate boxes for your analysis and click on “Submit” at the bottom of the page. Check “Error Analysis” only if you are processing a mock community sample, i.e. one made up of known species.

There are two features of the pipeline that I do not like. Chimera check is run separately on each sample using uchime in de novo mode. Representative sequences are given for each sample rather a single representative sequence for each cluster. I avoid these short-comings by using the command line tool.

Results are downloaded as a compressed file. The cluster folder contains a clust file and biom files containing only OTU tables, by default for distances from 0 to 0.5.

It is also possible to run steps of the pipeline separately using the links on FunGene Pipeline page.

You may try running the web-based FunGene Pipeline using the example data provided in MIE_nifH_ seqs_input.zip. Download that example data from here.