its_glue is a pipeline to batch process pairs of forward and reverse chromatagrams produced via Sanger sequencing. It wraps powerful tools into a simple workflow, and outputs collated, organised results. The pipeline covers basecalling, assembly of forward and reverse reads, ITS extraction, quality control, and OTU picking.
Check the project wiki for install guides for Linux, Windows Subsystem for Linux, and Mac, including installation of requirements. I recommend running the pipeline on Linux.
All scripts must be run from inside the top-level directory of the pipeline.
Trace to ITS:
The first section of the pipeline handles basecalling the input trace files using Tracy, extracting consensus sequences where possible, chimera detection and finally extraction of ITS regions using ITSx. Consensus sequences are preferred but where necessary ITSx will extract from single direction strands.
This functionality is all wrapped into a single script:
./pipe_one.trace_its.sh \
-t path/to/traces/directory \
-f fw_tag \
-r rev_tag \
-o path/to/output
The script expects pairs of forward and reverse traces as .ab1 or .scf, all in one folder (i.e. not in subfolders). It will attempt the process on every trace file it finds in the input directory. The path to this directory is specified by the -t flag.
-f and -r specify the tags used to identify forward and reverse reads in the filenames. They should be last before the extension and the filenames for fw and reverse reads should otherwise be identical. They cannot be -f or -r - those are reserved for the flags!
Specify an output directory with the optional -o flag - if not specified it will default to the current location, which you probably don't want
If you want to overwrite previous output (i.e. you are sending output to the same path as a previous run), use the optional -x flag.
Play with the script if you need to change parameters, but you shouldn't need to (unless you're trying to ID something other than fungi)
ITS to OTUs/Identifications:
The second section of the pipeline handles OTU picking. It first clusters sequences into OTUs at 97% similarity, using the single linkage clustering algorithm provided by vsearch. It then uses sintax, again implemented in vsearch, to assign taxonomy (as best it can) to each OTU using a reference database. I am personally dubious of all presently available methods of taxonomic assignment (e.g. see here and here), for a discussion of this implemenation see the wiki (TODO).
The functionality is wrapped into a single script:
./pipe_two.cluster_classify.sh \
-d path/to/reference/database \
-f path/to/input/fasta \
-o path/to/output
The script expects, as arguments, a reference database, specified by the -d flag, an input fasta, specified by the -f flag, and an output directory, specified by the -o flag. If you are using the first stage of the pipeline, the input fasta will be in the its directory of the output, and named ALL_ITS.fa.
See the project wiki for a guide to outputs for each pipeline stage (part-complete)
If you have issues with the pipeline please post them to the Issues tab of this GitHub repository. Please note that this is only for issues installing the pipeline itself or running it, not installing the dependencies. I'd love to have time to help people install dependencies etc. but I unfortunately don't - and they have their own GitHub repositories where you may ask questions.
If you have queries, comments, suggestions or questions about the pipeline please submit them to the Discussions tab. Hopefully this way we can build up something of a community FAQ.
I haven't really figured this bit out yet, but if you use this pipeline, or data produced from this pipeline, please credit me in some way - I suppose either through authorship (if appropriate) or citing (for now) this page. You should also be citing the appropriate publications for Tracy, ITSx, vsearch and sintax.