Fig 1.
Light micrographs of Labidocera madurae copepodite (A, B) and adult female (C,D). (A) Copepodite stage CIII, dorsal view (magnification: 4x). (B) Same copepodite as in A under fluorescent light showing expression of green fluorescent protein (GFP) (magnification 10x). (C) lateral view of the anterior portion of an adult female showing one dorsal and the ventral ocelli, feeding appendages and GFP expression (magnification 10x). (D) Lateral view of the same individual as in C under fluorescent light showing GFP expression at the base of the swimming legs (magnification 10x). Scale bar: 0.5 mm.
Table 1.
De novo assembly and annotation statistics.
Labidocera madurae RNA-Seq data from six samples were combined, quality filtered and trimmed and assembled using Trinity software [24].
Fig 2.
Diagram of the workflow used to generate the de novo transcriptome for Labidocera madurae and the three approaches used to test for completeness and quality of the assembly.
Fig 3.
Biological processes represented in L. madurae transcriptome.
Pie chart of the annotated transcripts including Gene Ontology (GO) terms belonging to the biological process (BP) category.
Table 2.
Comparison of de novo transcriptomes generated for non-model arthropods.
Table 3.
Four transcripts encoding “giant” proteins assembled using Trinity software in Labidocera madurae transcriptome. For each transcript, transcript length, predicted protein length, annotation name (NCBI), Accession No. of top blast hit (NCBI), E-value annotation (NCBI), protein family and protein function are listed.
Fig 4.
Labidocera madurae voltage-gated sodium channel sequences assembled by Trinity.
Diagram at top shows the four well-conserved domains (DI-DIV) bridged by less-well-conserved loops. Conserved domains are depicted vertically expanded to show approximate locations of six trans-membrane α-helical segments (colored bands labeled S1, S2-S6). Sodium-selectivity of the NaV1 transcripts (but not NaV2) is confirmed by the occurrence of four characteristic amino acids (aspartic acid, glutamic acid, lysine and alanine [DEKA]) in specific locations termed the "P-loops" [31]. Coverage by variants of three putative genes, Labma NaV1.1 Labma NaV1.2 and Labma NaV2 indicated by bars labeled with the i number assigned by Trinity. For Labma NaV1.1, no one sequence possessed all of the pieces (putative exons), so the overall span across the diagram represents a manual reconstruction generated by including all of the pieces from the different i’s. Gaps in sequences are indicated by fine dotted lines. Identical 5' (504 nucleotide) UTRs for i1-i7 have been omitted, as have the identical 3' UTRs (1518 nucleotides) of i1 and i2. Within each gene, corresponding residues across different i’s were identical (reflected in the same coloration of the bars) in almost all cases, except for the splice variant indicated in red for NaV1.1 i3. Sequences representing partial predicted proteins not initiated by an M at the N-terminal or terminated by a stop codon (“X” above the bar) at the C-terminal are indicated with a short diagonal bar. Positions of the domains for NaV2 differ somewhat from those of NaV1 shown in the top diagram and are indicated by thickening of the bars. Two sites of putative splice variation (Site I and II) are indicated below the NaV1.1 diagram, and one non-optional segment within Site I is designated "1" (96aa). Arrows in the NaV2 diagram indicate short optional pieces (gaps in the horizontal bars), and the overlap region between the two pairs of isoforms of 44 identical amino acids (aa) is indicated.
Table 4.
Labidocera madurae (Labma) voltage-gated sodium channel transcripts/predicted proteins.
Table 5.
Putative Labidocera madurae (Labma) circadian signaling system transcripts/proteins identified via in silico transcriptome mining.
Fig 5.
Alignment of five PDP1 protein sequences predicted from the L. madurae de novo transcriptome.
Four genes were predicted (I-IV). The first two sequences (Labma-PDP1-I-v1 and Labma-PDP1-I-v2) are likely to be splice variants, since they are identical except for a 9 amino acid long indel.
Fig 6.
Predicted gene mapping to the circadian rhythm pathway obtained through KEGG annotation.
Circadian rhythm pathway shown represents a map for Drosophila melanogaster (map04711). Highlighted boxes (green) represent L. madurae transcripts with coding regions (CDS) automatically annotated against the Kyoto Encyclopedia of Genes and Genomes (KEGG). PER, VRI, PDP1 were not identified by the automated annotation (white boxes).
Table 6.
Comparison across four possible reference transcriptomes generated from the de novo assembly for gene expression studies.
Reference transcriptomes—“Full”: complete de novo Trinity assembly; “Pred. genes”: retained a single (longest) isoform each Trinity-defined unique genes; “Full-CDS”: de novo Trinity assembly filtered using TransDecoder with only transcripts with predicted coding regions retained; “Pred. genes-CDS”: “Pred. genes” transcriptome filtered using TransDecoder with only transcripts with predicted coding regions retained. Number of transcripts, Bowtie mapping statistics and BUSCO analysis is given for each reference. Differential gene expression results include the number of transcripts that were included in the statistical analysis (expression level: > 1 cpm) and number of identified differentially expressed genes (DEGS) using either Bowtie or kallisto software as the mapping program.
Fig 7.
Non-proportional Venn diagram for the number of differentially expressed genes (DEGs) identified using four different transcriptomes as a reference for mapping of reads.
The references transcriptomes are defined as: “Full” with 211K transcripts (purple), “Pred. genes” consisting of longest transcript for Trinity predicted genes (yellow), “Pred.genes-CDS” consisting of transcripts with coding regions (CDS) from the “Pred.genes” (green) and “Full-CDS” consisting of transcripts with coding regions (CDS) from “Full” (pink). Relative transcript abundance as determined using kallisto, and DEGs were identified by statistical analysis using edgeR with P<0.05 and false discovery rate (FDR) cutoff at 5%.
Table 7.
Comparison among reference transcriptomes in the identification of differentially expressed genes (DEGs) between L. madurae copepodites and adult females among transcripts encoding for “giant” proteins, voltage-gated sodium channels and circadian system proteins.
Transcripts were identified as DEGs using a Benjamini-Hochberg corrected p-value <0.05.