Both datasets (this year's and last year's) were uploaded to UCL high performance compute cluster Legion and processed according to the bash scripts 2016_cr_join_no_golay.sh and 2017_cr_nojoin_no_golay.sh. The standard output from these scripts is saved in 2016_stdout and 2017_stdout, respectively.

The files contain several lines of codes starting with #, which are instructions for the HPC job queue. Such as how long the job will run on how many cores, how much RAM and storage is required.

For 2016's data both reads were joined, prior to de-multiplexing. As 2017's second read data was of lower quality, joining was skipped and only read1 was processed.

After picking OTUs with the Greengenes reference the BIOM-formated tables were merged with the command in merge_otu_tables.sh.

Any of the phylogentic trees found in 2016/otus or 2017/otus can be used for further analysis (e.g. the core_diversity_analyses.py script), as this is the Greengenes reference tree.