40% of reads survived the pipeline

[dsamoht]

Hi, We have difficulties to pinpoint the reason why we loose ~60% of our sequences after the dada2 pipeline. We ran the pipeline multiple times with a range of parameters for maxEE and truncLen. Here are some results of our "best" run.

Reads tracker (maxEE=c(2,5), truncLen=c(200, 190), other parameters at default) : reads_tracker_2_5_200_190.csv

Error plot (FWD):

Error plot (REV):

run characteristics:

• Environmental water data ;
• V4-V5 region of the 16S (515FB: 5’-GTGYCAGCMGCCGCGGTAA-3’, 926R: 5’-CCGYCAATTYMTTTRAGTTT-3’);
• Primers have been removed with cutadapt (with --discard-untrimmed);
• Miseq v2 (250PE);

FASTQC plots: FWD: REV:

We would greatly appreciate your opinion on this problem, thank you!