Metatranscriptomic and transcriptomic databases (DB4S) of Spodoptera frugiperda larvae guts from Northern Argentina (Tucumán province)

Abstract

Spodoptera frugiperda is a noctuid moth that devastates various crops including corn, rice and cotton, and is found in most of the American continent. The purpose of this study was to integrate gene expression data from S. frugiperda guts and their associated metatranscriptomes, under natural and controlled conditions. For this, four S. frugiperda samples from the province of Tucumán (Argentina; subtropical region) were analysed. Specimens were obtained from different environments, altitudes and food sources, namely: 1) a transgenic maize (Zea mays) field at 495 m.a.s.l. where insecticides and fertilisers were applied (named MM; 26o49’50”S; 65o16’59.4”W); 2) Sorghum halepense at 495 m.a.s.l. (MS; 26o49’50”S; 65o16’59.4”W); 3) a maize field at 2283 m.a.s.l. where no insecticides or fertilisers were used (TV; 26o55’40.75”S; 65o45’19.90”W) ; and 4) a colony established from larvae originally collected from the same transgenic maize field as Sf_MM, reared for 9 generations under controlled conditions on an artificial diet adapted from [8], without the addition of antibiotics (BT). For all samples, total RNA extracted from fifth instar larvae guts (two digestive tracts per sample), was submitted to a modified one-step reverse transcription and polymerase chain reaction sequence-independent amplification procedure, as described previously. High-throughput pyrosequencing of the samples was performed using a Roche GS FLX (Macrogen Inc., Korea), yielding ~1Gb of metatranscriptomic reads with lengths of 50 to 1600 bases (nt) (652 nt average). Raw sequence reads were trimmed to remove nucleotides derived from the amplification primers using a custom application. Below follows an outline of the main steps we followed to create the uploaded databases: I.Sequences were compared locally to a combined nucleotide database (nt16SLep = “Non-redundant” nucleotide sequence (nt) database + 16S rRNA gene (16S) database + Lepidopteran whole genome shotgun (Lep) projects completed at the time of the analysis) using BLASTN (Altschul et al., 1990) with a 1e-50 cutoff E-value, and to the protein database (nr = non-redundant protein sequence) using Diamond (Buchfink et al., 2014) with a 1e-17 cutoff E-value. II.The homology search results were then processed as follows: Step A: The output files from both homology searches were processed with MEGAN, a software which performs taxonomic binning and assigns sequences to taxa using the Lowest Common Ancestor (LCA)-assignment algorithm (Huson et al., 2007). Taxonomic and functional assignments performed by MEGAN for each sequence were then exported using a MEGAN functionality. Note: MEGAN computes a “species profile” by finding the lowest node in the NCBI taxonomy that encompasses the set of hit taxa and assigns the sequence to the taxon represented by that lowest node. With this approach, every sequence is assigned to some taxon; if the sequence aligns very specifically only to a single taxon, then it is assigned to that taxon; the less specifically a sequence hits taxa, the higher up in the taxonomy it is placed. Step B: The output files from both homology searches were also processed with a custom bash script. This script parses the homology search output files and generates two files (one for each homology search) containing the name of each sequence, its best hit (or no hit) and the corresponding E-value. III. Create local database (Step C): All this information (from the exported MEGAN files and from the bash script output files) was then used to create a local SQLite database which included all the available information for each sequence (from both homology searches).