Comments on revised manuscript

I have now carefully read the revised version of this manuscript and I am happy with the changes that the authors implemented as a response to my comments and the comments of the other reviewer. The paper is now much more clear, especially in the methodological section and the limitations of the use of the novel sequencing platforms/formats is sufficiently discussed. Minor comments that should be made in the present paper: L58: change "bacteria" to "bacterial" L65-66: the last part of this long sentence is difficult to comprehend and should be rephreased. I suggest to divide the long sentence into two L68-69: change "produces" to "produced" L84: delete "in" L98: please explain the abbreviation "ONT", likely "Oxford Nanopore Technologies" L162: the detail of the amplification methods should be expanded at least stating the primer pairs (names and sequences) used and targeted molecular markers; from the text it appears as if ITS2 was the marker selected, yet lines 361 and 366 discuss length differences in ITS1 L246: replace "common fungi several species" with "common fungal species" L248-251: the misclassification of fungal taxa was not due to the bad performance of the sequencing platform, it was because of the low variability of the ITS2 marker. I suggest to change the text to state that genus level assignment was reached for these taxa since multiple species had the same ITS2 sequence L264-265: the main reason is that the PCR bias (preferential PCR amplification of certain templates) skews the representation of taxa if the DNA is mixed prior to amplification L331-346: this section is unclear; it should be specified which primers (primer names and sequences) with what barcodes were used for each conditions; if different primer pairs were used for different sequencing platforms, it is unclear what is the use of this comparison. This should be either clarified and explained all this section may be removed. L381: delete "so" L387-392: I suggest that this part is either removed or it is clearly described why the authors are sure that PCR replicates are not necessary (which is against all present recommendations). While the increasing fidelity of polymerases can be a fact, the main problems with parallel PCR is not errors (due to low fidelity) but random effects where primers align to templates with random frequencies. This statistical effect is impossible to handle by increasing polymerase fidelity while it is easily handled by PCR replication. L424-426: This statement is rather obvious, I suggest to delete it.