mothur v.1.31.0

We are very happy to announce the release of mothur v.1.31.0 and have a number of exciting new features to tell you about. First, we have re-implemented the SparCC program first described by Friedman and Alm. It calculates correlation coefficients that are not sensitive to the hideous effects of relative abundance data. You can run it with the sparcc command. Second, we have implemented the approach described by Holmes and colleagues for assigning samples to metacommunities using Dirichlet multinomial mixture models that we are calling get.metacommunity. Think of it as a robust way of assigning samples to bins for doing things like finding enterotypes. Third, through a project carried out with Google’s Summer of Code we have created the classify.shared command that currently implements a feature selection algorithm using the random forest algorithm. It’s faster and more accurate than the more commonly used R packages. Finally, since people are staring to develop their own mock communities and are interested in using shhh.flows with Ion Torrent data and new 454 platforms, we have created make.lookup for making your own lookup file to use with shhh.flows. In our testing, we don’t see much of an effect, but hey, you might. As always, we have a number of new features that we’ve added to existing functions. As always, we’ve added a number of new options to make your life easier.

Finally, a couple of important announcements. First, Pat will be holding an R workshop in July and a mothur workshop in August. We assume no experience for either workshop and you can get more information by emailing Pat by clicking on the link at the top of this page. Second, we’d be lying if we told you we didn’t keep track of how many times the original mothur paper has been cited. Watching the number tick up has been humbling over the past 3 or so years. Well, we’re currently sitting at 983 citations. Thank you, thank you, thank you! We’ll be watching Web of Knowledge to see who gets us to 1000 and will be sending them a special gift in appreciation. We’re honored that we’ve been able to help so many people do great research!

New commands

  • get.metacommunity - assign samples to metacommunities using the algorithm of Holmes and colleagues
  • sparcc - calculate the correlation between OTUs using the algorithm of Friedman and Alm
  • classify.shared - identify features (i.e. OTUs) that allow one to distinguish between sets of communities using the random forest machine learning algorithm.
  • make.lookup - create custom lookup files for use with shhh.flows

Feature updates

Bug fixes

  • added OTUlabels to shared file in remove.rare - fixed in 1.30.1
  • windows paralellization of pairwise.seqs cutoff was not passed to thread - fixed in 1.30.1
  • sffinfo - clipQualRight=0 caused issues with rightIndex value in flow file.
  • pcr.seqs - For forward primer trimming with aligned sequences and keepdots=t. If the character before the first primer base was a base and not a gap the base was not trimmed. https://forum.mothur.org/viewtopic.php?f=4&t=2209 - fixed 1.30.2
  • summary.seqs - 32bit machines processing a file larger than 4G could hang.
  • pcr.seqs - removing primers from aligned sequences with keepdots=f could result in an unaligned dataset if primers were not found at the same locations. Added gaps to preserve alignment.
  • classify.tree - when name file is given everything is unclassified - https://forum.mothur.org/viewtopic.php?f=3&t=2161&p=6116#p6116
  • remove.groups(groups=notValidGroupName, ...) mothur removes all of the groups. Fix will ignore invalid group and continue.
  • sff.multiple - fixed order parameter to use A, B and I.
  • cluster.split - when splitting by taxonomy mothur was saving temp fasta files as the “current” fasta file.
  • indicator - fixed bug with pvalues when multiple processors were used.

Changes to wiki

Registered users

2542