Vasco Elbrecht

20% of reads get discarded in PE merging with Vsearch

2

alignment score too low, or score drop too high, vsearch v2.19.0_macos_aarch64 Need to check setting changes

need to do a few tests here

Cluster ESVs to reduce table size and be better than OTU clustering allone = )

cat command can't handle more than 1000 files at once. Temporally fixed, for up to 2000 files.

Relevant for servers / large datasets https://www.r-bloggers.com/long-running-r-commands-unix-screen-nohup-and-r/

> usearch -fastx_subsample "../D_Minmax/_data/GMP-04606)CCDB-S5-0053)CBGMB-00003_cut_minmax.fasta" -fastaout "_data/GMP-04606)CCDB-S5-0053)CBGMB-00003_cut_minmax_N50000.fasta" -sample_size 50000 -sizein -sizeout Solution, count reads first and then copy the file over if it has fewer reads than should be subsetted

implement toupper or different fasta read in method. Thank you Evgeny Shcherbakov for finding this issue.

based on table 7 in https://esajournals.onlinelibrary.wiley.com/doi/full/10.1002/ecs2.3193