Figure 1.
Schematic overview of the library preparation process.
Both RAD-tag (left) and whole-genome shotgun (right) library preparation methods start the same way, and diverge only at the final stage. (a) DNA is heated to denature the template strands. (b) Terminal deoxynucleotidyl transferase (TdT) is used to add a riboguanidine tail of a determined length [44]. (c) Priming with the Illumina P2 adaptor sequence, the Klenow exo- fragment generates the second strand. At this point, T4 DNA polymerase treatment is necessary to blunt the DNA fragments. After (d’) for RAD-tag sequencing, EcoRI is used to digest a subset of the fragments. (d’’ and e) a final ligation step adds the P1 Illumina adaptor sequence. Barcodes are ligated in-line, upstream of the read one sequencing primer binding site. After ligation of the final adaptor sequence, fragments are PCR-amplified to complete the sequencing adaptor. All libraries contained in-line barcodes in front of the read one sequencing site.
Figure 2.
Fragment sizes of extracted DNA for ants (A) and fruit flies (B).
Virtually all fragments in the libraries were less than 100-axis are derived from Bioanalyzer traces, and are somewhat arbitrary, giving only an approximate indication of extraction yield. See Table 1 for more details on the actual yield from each extraction. We failed to amplify an approximately 700-bp fragment of mitochondrial DNA from these specimens using a popular primer set [22].
Table 1.
Specimens used in this study, and DNA yields.
Figure 3.
Images of representative specimens before and after extraction.
Specimens before extraction are in first and third rows, and after extraction in second and fourth rows. For ant photos, the same specimen is depicted before and after extraction. For Drosophila, different specimens from the same collection series were used for the before-and-after comparison. Specimen damage to ant was minimal, consisting only of eye de-pigmentation. The more fragile Drosophila specimens were more greatly affected, and their eyes show signs of partial collapse. The more fragile fruit fly specimens also showed signs of greater mechanical damage due to handling.
Figure 4.
Bayesian consensus cladogram of Hawaiian fruit flies.
The topology is consistent with molecular and karyotypic studies [40]–[42], and all nodes were resolved with 1.0 posterior probability. None of the previous molecular studies could resolve all of the nodes. Wing photographs from [45], except D. paucipuncta, which is from the present study.