Figure 1

Transcriptome assembly workflow and alignment to D. rerio reference proteome. A. RNA was extracted from brain and spinal cord tissue of adult A. leptorhynchus, fragmented and barcoded, and strand-specific cDNA libraries were created for Illumina sequencing. Reads were trimmed using several strategies and then normalized in silico prior to de novo assembly with Trinity. Transcript reconstruction was performed using several strategies and then benchmarked using BLAST to maximize transcripts aligning to a D. rerio reference proteome. The best assembly was then further annotated. B. Out of the transcripts from the entire A. leptorhynchus assembly with any alignment to a D. rerio reference protein, most transcripts aligned to <40% of the reference D. rerio proteins. C. When filtering the assembly for transcripts with FPKM ≥ 1, the relative proportion of transcripts with less than complete alignments was reduced. D. When selecting the assembly transcript with the longest D. rerio alignment within a given contig, this greatly reduced the relative proportion of less than complete alignments. E. This distribution was preserved when considering only transcripts with FPKM ≥ 1. F. Out of the entire assembly, more than half of the transcripts had at least some alignment to a reference D. rerio sequence. Similarly, more than half of the sequences with FPKM ≥ 1 aligned to a D. rerio protein sequence. G. Using only transcripts with FPKM ≥ 1, nearly 60% of the reference D. rerio proteome had at least one assembly transcript with an alignment (24,112 sequences), which represented approximately 80% of the D. rerio sequences that were hit by the entire assembly (30,121 unique sequences).