Truseq fragment library

To provide a reference assembly for mapping reads from the ddRAD data, we generated libraries that were optimized to be assembled using the ALLPATHS‐LG assembler (Gnerre et al., 2011). We used flies collected from the rearing colony maintained at the Petapa San Miguel Mexfly Mass Rearing Facility in Guatemala, collected from the Family 10 black pupae genetic sexing strain, and extracted DNA as for the ddRAD specimens. We constructed a 180‐bp insert Illumina TruSeq fragment library from 500 ng DNA from a single female individual, and an Illumina Nextera mate‐pair library targeting a 3‐kb insert size from a pool of sibling male individuals (eight individuals in total), to have sufficient DNA for mate‐pair library construction. The fragment and mate‐pair libraries were each sequenced on a lane of Illumina HiSeq4000 with 2 × 100 bp paired‐end sequencing and 2 × 50 bp paired‐end sequencing, respectively. We constructed a scaffold assembly from raw reads of both libraries using ALLPATHS‐LG (Gnerre et al., 2011). We performed k‐mer‐based error correction (using the ALLPATHS‐LG pipeline) to the fragment library and then ran the pipeline with default settings except for the addition of the “HAPLOIDIFY = TRUE” parameter.

Library preparation methods were used to optimize genome assembly with the ALLPATHS-LG assembler. A 180-bp insert Illumina TruSeq fragment library was constructed from 500 ng DNA extracted from a single GSS male. This individual was the F2 offspring of an isolated mating between two GSS parents. Additionally, two Illumina Nextera mate-pair libraries targeting a 3- and 8-kb insert size, respectively, were constructed using DNA from a pool of sibling GSS males that were derived from the same isolated mating parents as the individual in the fragment library. The fragment and mate-paired libraries were sequenced using 2 × 100 bp sequencing on the Illumina HiSeq 2500 in High Output mode. The SRA accessions for each library, along with additional read counts and approximate read depths, are presented in Table 1. Raw reads from the fragment and mate pair libraries were used to construct a scaffold assembly using ALLPATHS-LG (v.44837) (Gnerre et al. 2011 (link); Ribeiro et al. 2012 (link)) with default parameters, with the exception of addition of “HAPLOIDIFY = TRUE.” Kmer-based error correction of the fragment library was performed prior to assembly as part of the ALLPATHS-LG pipeline. The draft scaffold assembly was integrated with linkage data (described in more detail in Linkage mapping and QTL analysis) and placed into chromosome-scale superscaffolds.

The high levels of gene order synteny found amongst the Lepidoptera (Pringle et al., 2007 (link)) allows markers to be developed to cover all 31 chromosomes based on their location and non-paralogous nature in Biston betularia, Melitaea cinxia—species with 31 chromosome pairs and for which linkage maps are available, and also Bombyx mori with 28 chromosomes pairs (Ahola et al., 2014 (link); Duan et al., 2010 (link); Van’t Hof et al., 2013 (link)). The coding sequence of identified genes acquired from the B. mori database (SilkDB) was blasted against a local database to provide H. bolina-specific sequence. To create the local database, DNA extracted from a Thai female butterfly was used to create two Illumina paired end libraries (TruSeq fragment library and Nextera large-insert library), which were sequenced on the Illumina HiSeq platform. Reads were assembled using CLC Genomics Server to create a draft de novo assembly, which was used to create a custom, local nr BLAST database in Geneious Pro v. 5.6.6. Retrieved H. bolina sequence was aligned to B. mori coding sequence to infer intron/exon boundaries, so that primers could be designed to amplify intronic regions to develop informative markers, which were then sequenced through the Sanger method. Marker information is detailed in Table S1.