For insecticide treatment experiments, adults of two MED populations, PyriR, which is susceptible to the insecticide Mospilan (acetamiprid), and 9-2103, which is resistant, were fed on cotton seedlings (Gossypium hirsutum L. cv. Acala) treated with the insecticide Mospilan at an LC30 dose (lethal concentration required to kill 30% of the population; 2 ppm for PyriR and 100 ppm for 9-2013) with the dipping method, as previously described [64 (link)]. Whiteflies fed on untreated cotton seedlings were used as controls. The experiments were conducted under standard rearing room conditions of 25 °C, 50% relative humidity, and a light regime of 10 h light and 14 h dark. Three to four biological replicates, each containing a pool of 200–500 adult whiteflies, were collected from each treatment. The insects were kept at –80 °C until use.
Total RNA was purified using the Ambion TRIzol Reagent (Thermo Fisher, USA) according to the manufacturer’s protocol. Strand-specific RNA-Seq libraries were constructed following the protocol described in Zhong et al. [65 (link)] and sequenced on the Illumina HiSeq 2500 system. Raw RNA-Seq reads were first processed to remove adapter and low-quality sequences using Trimmomatic [57 (link)]. Reads shorter than 40 bp after trimming were discarded. The resulting reads were then aligned to the ribosomal RNA database [66 (link)] and the three bacterial symbiont genomes using Bowtie [67 (link)], allowing up to three mismatches. The aligned reads were not used for further analysis. To assist gene prediction, the high-quality cleaned RNA-Seq reads were aligned to the assembled B. tabaci genome using TopHat [68 (link)], and the aligned reads were assembled into transcripts using Cufflinks [69 (link)]. For gene expression analysis, the RNA-Seq reads were aligned to the assembled B. tabaci genome using HISAT [70 (link)]. Raw counts for each B. tabaci predicted gene were derived from the read alignments and normalized to fragments per kilobase of exon model per million mapped fragments (FPKM). Differential expression analyses were performed using edgeR [71 (link)]. The resulting raw P values were adjusted for multiple testing using the false discovery rate (FDR) [72 ]. For each comparison, genes with FDR <0.05 and fold change no less than 1.5 were considered as differentially expressed genes.