Heliconius hecalesia, Hewitson, 1854

Brower, Andrew V. Z., 2018, Alternative facts: a reconsideration of putatively natural interspecific hybrid specimens in the genus Heliconius (Lepidoptera: Nymphalidae), Zootaxa 4499 (1), pp. 1-87 : 75-79

Putative H. hecalesia Hewitson, 1854 hybrids

Identity: 1.0

Authenticity: 0.1

Overall reliability: 0.1

Identity: 0.5

Authenticity: 1.0

Overall reliability: 0.5

Although Genbank accession numbers were not reported in Dasmahapatra et al. (2007), most (but not all) of the sequences referred to in the paper are available in Genbank. Close examination of sequences from the "hybrid" and other Heliconius revealed some interesting facts. As reported, 06-921's 2119 bp mtDNA sequence (Genbank accession code AM709828 ) is virtually identical to that of H. ethilla aerotome 02-975 ( AM709826 ), which strongly suggests that the mother of the specimen was an H. ethilla . The nuclear genes, with two alleles each, are more complicated. Teasing apart a chromatogram with heterozyous sites into two separate alleles is not an easy feat, but other than some description of "deconvolution" of sequences of variable length, Dasmahapatra et al. (2007) did not describe how they determined the sequences of alternate alleles, other than by comparing the heterozygous sites to the sequences of the two putative parental species. Needless to say, calling the bases so that they match one or the other parental sequence is hardly an independent corroboration of the allelic similarity of the "hybrid" to the parents. Indeed, several of these genes are known to exhibit dramatic heterozygosity of intron sequence and length within "pure" (i. e., not hybrid) individuals ( Brower 2011), yet Dasmahapatra et al. (2007) used single sequences of H. ethilla and H. melpomene in their Neighbor-Joining analyses, apparently assuming that every other specimen except 06-921 was homozygous.

The two 06-921 Mpi "alleles" ( AM709819 , AM709820 ) were short (315 and 313 bp, respectively), and differed from one another only by a 2 pase-pair indel. The former was identical to an Mpi sequence from H. ethilla aerotome 02-975 (AM709815.1), while the latter was identical to an Mpi sequence from H. melpomene amaryllis JM1917_A (AY332454.1) and from several other H. melpomene races. However, some H. ethilla Mpi sequences are more similar to H. melpomene sequences than they are to one another, and some Mpi sequences included in Dasmahapatra et al. (2007) fig. 2 are not present in GenBank. These problems cast doubt upon the value of these sequences as evidence for interspecific hybridization (cf. Brower 2011).

The two 06-921 invected sequences ( AM709837 and AM709838 ), aligned below, were only 52 and 53 bp long, respectively:

CTTTTGTATCTTTTTTGTTTT-ATTCAAATTACAAAGTTTGTAATACATACAT ATTTTGTATCTTTTTTGTTTTTATTCAAATTATTAAGTTTGTAAAGGCTCTTA

Most Heliconius invected sequences in Genbank are over 400 bp long, and Dasmahapatra et al. offered no explanation why these sequences are so truncated. Using Genbank's BLAST query (https://blast.ncbi.nlm.nih.gov/ Blast.cgi), the former is a perfect match to H. ethilla aerotome 02-3 sequence ( AM709835 , also only 52 bases long), while the latter is identical to a number of H. melpomene and H. heurippa sequences. However, when taken in a broader context (e. g., Brower 2011), neither available alleles for H. melpomene nor for silvaniform taxa form coherent groups for this gene. In any event, these tiny sequences do not provide much evidence for any pattern at all. Note also that the last 9 bases of these two sequences are not parsimoniously viewed as homologous sites when aligned with longer invected sequences from other Heliconius .

The specimen's two Tpi sequences, AM709812 (563 bp) and AM709813 (411 bp) differ from one another primarily by the presence or absence of several long indels. The longer sequence is identical to three H. ethilla sequences, while the shorter one is similar to H. melpomene sequences. As with Mpi, this gene region is known to exhibit individual heterozygosity in these indel regions ( Brower 2011), and it is certainly not a safe assumption to use a single allele from local H. ethilla and H. melpomene races to authoritatively assign the affinities of alternate alleles from the "hybrid."

There is only one Tektin sequence in Genbank for 06-921 ( AM709690 ), and Dasmahapatra et al. (2007) apparently sorted ambiguous chromatogram peaks from this sequence to match either an H. melpomene or an H. ethilla sequence. As noted above, such a procedure begs the question of the specimen's identity, and does not constitute evidence. The fifth nuclear gene, Rpl5, was uninformative by Dasmahapatra et al. 's admission, due to non-monophyly of alleles of H. ethilla and H. melpomene with respect to one another.

To sum up, Dasmahapatra et al.'s presentation of the molecular evidence endeavored to show that 06-921 is a H. ethilla x H. melpomene F1. However, the only marker that is not compromised by potentially unrealistic simplifying assumptions is the mtDNA, which only tells us about one parent. The molecular evidence thus leaves room for doubt about the plausibility of this specimen's hybrid origin.

Identity: 0.75

Authenticity: 0.9 (vague details of when and by whom the specimen was collected).

Overall reliability: 0.68