On 2015 Sep 25, Kenneth Witwer commented:

As stated in a letter to the editor (Witwer KW, 2015), monocot-specific MIR528 was the apparently most abundant and best-absorbed miRNA in this study, more abundant than all other detected plant miRNAs combined. Watermelon--the only material ingested by the study volunteers--is a dicot. No sequences identical to mature or precursor MIR528 are found in watermelon sequences in public databases, nor in any currently available dicot genomes. In response to this observation and as evidence for the existence of dicot MIR528, the authors refer to Lin Y, 2013, in which a putative MIR528 relative was identified in an RNA sequencing library prepared from the dicot Dimocarpus longan. This sequence, CTGGAAGTGGATGCAGAGGG, has no fewer than five nucleotide differences from monocot MIR528, gUGGAAGGGGCAUGCAGAGGAGc (lower case letters are precursor nts surrounding the mature miRNA). No such sequence appears to be found in the public genomes or transcriptomes of watermelon or other dicots. The putative longan sequence matches better to various sequences with a common 18-nt stretch found in dicots, but also in animals, than to MIR528. Even if the putative dicot MIR528 sequence were a microRNA, the two differences from monocot at the 3' end would interfere with stem-loop reverse transcription by the monocot-specific assay used by the authors. In any case, for a dicot plant to express an otherwise monocot-specific miRNA, the sequence would first have to be present in the genome of the dicot. Unless and until genomic evidence is provided, MIR528 detected in watermelon or watermelon-fed humans must be presumed to be a contaminant or other artifact.

On 2016 Mar 08, Kenneth Witwer commented:

Because of continuing interest in this topic, an extended version of these comments has been deposited on the "Negative Results" track on ResearchGate.

On 2015 Sep 25, Kenneth Witwer commented:

As stated in a letter to the editor (Witwer KW, 2015), monocot-specific MIR528 was the apparently most abundant and best-absorbed miRNA in this study, more abundant than all other detected plant miRNAs combined. Watermelon--the only material ingested by the study volunteers--is a dicot. No sequences identical to mature or precursor MIR528 are found in watermelon sequences in public databases, nor in any currently available dicot genomes. In response to this observation and as evidence for the existence of dicot MIR528, the authors refer to Lin Y, 2013, in which a putative MIR528 relative was identified in an RNA sequencing library prepared from the dicot Dimocarpus longan. This sequence, CTGGAAGTGGATGCAGAGGG, has no fewer than five nucleotide differences from monocot MIR528, gUGGAAGGGGCAUGCAGAGGAGc (lower case letters are precursor nts surrounding the mature miRNA). No such sequence appears to be found in the public genomes or transcriptomes of watermelon or other dicots. The putative longan sequence matches better to various sequences with a common 18-nt stretch found in dicots, but also in animals, than to MIR528. Even if the putative dicot MIR528 sequence were a microRNA, the two differences from monocot at the 3' end would interfere with stem-loop reverse transcription by the monocot-specific assay used by the authors. In any case, for a dicot plant to express an otherwise monocot-specific miRNA, the sequence would first have to be present in the genome of the dicot. Unless and until genomic evidence is provided, MIR528 detected in watermelon or watermelon-fed humans must be presumed to be a contaminant or other artifact.