OBITools bioinformatics workflow
Sequences were aligned using illuminapairedend with a minimum quality score of 40. Then, sequences were demultiplexed using ngsfilter and obisplit. Sequences with ambiguities were removed using obigrep, and filtered using obiclean with a threshold of 0.05 before the taxonomical assignation using ecotag. Further bioinformatic filters were applied to remove PCR or sequencing related errors and non-specific amplifications: i) removal of amplicons with less than 10 reads per PCR, ii) removal of the non-specific amplifications (non-fish orders), iii) removal of the amplicons which size was not comprised in the range of the targeted sequence (50 - 75 bp) and iv) cross-sample contamination cleaning removing amplicons with less than 1/1000 reads per PCR run (i.e. tag jumps, (S1) Schnell et al 2015) and in less than 2 PCR runs from the same sample ((S2) Ficetola et al 2015). An additional check was performed to remove species from contamination using DNA extraction and PCR controls. For instance, 34 to 230,685 reads in 7 sites passed through the bioinformatics filters and were assigned with a similarity of 98.4% and 100% to the Atlantic salmon (Salmo salar ), a species absent from Indonesia but which DNA was reported as a contaminant in some marine eDNA studies (e.g. (S3) Thomsen et al 2016).