Lessons Learned: the Importance of Biological Curation.

REPLY

Proper species identification of sequenced fungal isolates or strains is imperative for the interpretation of genomics data. Through this letter (1), and through another recently published research note (2), it has come to our attention that one of the isolates presented in our previous publication (3) was misidentified. The presumptive Aspergillus parasiticus isolate NRRL 2999 has been shown to be an Aspergillus flavus isolate, a clonal derivative of A. flavus NRRL 3357 (2). In our recent publication presenting two new reference genomes for A. flavus (4), we performed a phylogenomics analysis of several publicly available Aspergillus assemblies and noted the sequence similarity between A. flavus NRRL 3357 and A. parasiticus NRRL 2999. We also noted a close relationship between A. flavus WRRL 1519 and A. oryzae RIB40 and that A. oryzae RIB40 grouped among A. flavus isolates, not as an outgroup. Based on these analyses, we concluded that the A. flavus species is polyphyletic. The revelation that NRRL 2999 is indeed an A. flavus isolate clonal to NRRL 3357 (2) does point toward A. flavus as a distinct species from A. parasiticus but does not change the conclusion of the potential polyphyletic nature of A. flavus based on the position of A. oryzae in the analysis (4). All additional conclusions were based on phylogenetic relationships defined within the scope of the manuscript (4) and were therefore independent of species labels.

As reported by Smith et al. (5) on “the early scientific literature,” the issue of isolate identification is further complicated for aflatoxigenic fungal isolation and classification. Over the years, there have been occurrences of the same strain being assigned many different designations by different research groups. One interesting example is NRRL 2999, which was originally isolated from Ugandan peanuts in 1961 (6, 7). Over time, this isolate has received numerous other designations, such as Austwick strain V. 3734/10, Hodges M-3, SYS-4, ATCC 56775, ATCC 26692, CMI 91019b, NRRL 5862, ATCC 15517, SU-1, and SRRC 143, depending on the research groups the isolate was received from. It is now accessioned as IMI 91019b, NRRL 2999, and ATCC 15517, serving as the type strain for the American Type Culture Collection (2). There are two genome sequence data sets for A. parasiticus SU-1 (8) in NCBI. Since A. parasiticus SU-1 and NRRL 2999 are actually the same isolate, we plan to remove the current sequence data for the misidentified “NRRL 2999” from NCBI (GenBank accession no. CP051027 to CP051034) (3).

This situation does highlight the danger inherent in large-scale genome sequencing experiments, where the identity of isolates is presumed correct based on identification provided by the source of an isolate. Therefore, we must join with Houbraken et al. (1) in agreement that additional safeguards should be taken to ensure that the proper identification of isolates be determined before conclusions are made based on sequencing data.