The authors have not corrected their grossly mis-interpreted presentation of our methods. Specifically, we communicated this: "This figure shows the authors have grossly mis-represented our methods, which impact their major conclusions of the paper and mis-represent our work. Specifically, Babaian and Edgar suggested in their Fig.3 (“Defining RdRp boundaries for sequence-based classification”) that our approach to design vOTUs is wrong because we use fragmented RdRp sequences. This is not true. Just as the Wolf et al precedent, we only used complete RdRp sequences for defining OTUs as we also felt that was most appropriate. As we have sought to clarify with these authors on multiple occasions:

1. We used at least 90% complete RdRp domain sequences for classification purposes.
2. Though we used benchmarked (figure S9) genome-based definition for vOTU for ecological analyses (benchmarked against complete RdRp domain sequences), the range of agreement between the complete-RdRp- and genome-based definitions is 93% (for the shortest contigs in our data; i.e. 1kb) to 99% (for contigs >10kb). The "fragments” Babaian & Edgar (2022) show in Fig.3 do not represent either of these cases, and hence, are misleading for both taxonomy and vOTU definitions, and misrepresent our published work." In the correction the authors have made they have cherry-picked a fragmented sequence from the GenBank datasets, not even our datasets to make their case. Edgar in particular has consistently mis-represented our work here, as well as in follow-on efforts to publish that all grossly mis-represent our methods and our results time and time again. I do not know how else to stop this pattern of deliberate mis-information and confusion.