Ler_transcriptome.fa

Using bycatch of commercial fisheries, samples from embryos undergoing paired fin development were collected from a wild population of cownose rays (<i>Rhinoptera bonasus; Rbo</i>) in Chesapeake Bay. RNA extractions were made from various fin domains and sequenced on an Illumina HiSeq4000 and Illumina MiSeq. Adapters and low quality sequences from cownose ray data were trimmed from fastq files using the TrimGalore wrapper for Cutadapt and FastQC. Overlapping reads were combined with Flash prior to assembly and orphaned reads were retained. To reduce the impact of genetic variability, data from fin domains of only three individuals were used in the assembly, including: two regions of a stage one embryo (broad anterior pectoral fin region and mid-anterior pectoral fin), one region of a stage seven embryo (clasper) and eight regions of stage three embryos (cephalic lobe, anterior pectoral fin, mid-anterior pectoral fin, mid-pectoral fin, mid-posterior pectoral fin, posterior pectoral fin, zipper region, and pelvic fin). Sequencing with both HiSeq (stage 3 samples, 100bp PE reads) and MiSeq (stage 1 and stage 7 samples, 300 bp PE reads) was conducted to help generate full length transcripts. Abyss was used to assemble contiguous sequences (contigs) from the paired end (PE) and orphaned sequence data with 10 kmer sizes ranging from 25-70 nucleotides. Contigs of 500 nucleotides or greater were summarized according to similarity (e-value &lt; 1e-160) with BLAST+ tools. <br>Using iterative BLAST searches and pairwise alignments, contig sequences with strong e-values (&lt;1e-10) were annotated and used in the final transcriptome. Specifically, initial screening and annotation of contigs was conducted iteratively by best-reciprocal-blast using peptide, coding sequence, and non-coding libraries from Danio rerio (GRCz10) with BLAST+ tools and a minimum e-value threshold of 0.1. Each R. bonasus contig was paired with D. rerio sequences and the lowest average e-value was chosen as the annotation. Therefore, more than one R. bonasus contig could be designated with the same D. rerio sequence, allowing for potential redundancy in the de novo R. bonasus transcriptome assembly. The longest open reading frames (ORFs) that were unannotated to D. rerio sequences were annotated by BLASTp using the RefSeq vertebrate protein database (downloaded 4-4-2017). Remaining unannotated contigs were designated by their top forward BLAST hit in D. rerio. For consistency, the same process was used for the little skate (<i>Leucoraja erinacea; Ler</i>) using publicly available stage 31 <i>L. erinacea</i> 100bp PE RNA-seq data (SRA, PRJNA288370)<br>