Since some transcripts could be assigned multiple KOG functions, altogether 46,291 functional annotations were produced and all hit transcripts were grouped in 25 categories (Tables 3 and 5).
In both libraries, about 50% of the contigs assembled did not have a match in either database (designated "no hit" transcripts).
The no-hit transcript that is encoded on H3_4A11 (Mex_Nohits_2574_Contig_1) is significantly downregulated during spinal cord regeneration, while the no-hit transcript that is encoded on H3_61C19 (Mex_Nohits_221_Contig_2) is significantly upregulated during spinal cord regeneration and downregulated during skin metamorphosis [ 23, 24].
The top-hit transcripts were selected as unigenes.
These novel genes correspond to the no-hit transcripts described above.
Although some no-hit ESTs are truncated versions of known genes, it is possible that many of the ~2000 no-hit transcripts in the Ambystoma EST database correspond to novel axolotl genes.
In parallel, BLASTX search with the SwissProt set of UniProt protein sequences identified significant similarity with known proteins for 4,023 transcripts (20.5%) with a total of 62,630 hits, with 16 hits per transcript on average.
A total 220,388,641 and 202,416,138 reads were generated from Vero Beach and Gainesville populations and a total of 88,402 transcripts hit on annotated transcript sequences of Cx. quinquefasciatus (Table 2).
The proportions of similarity hits were somewhat lower than those hit the transcript database (Additional file 11).
For shotgun transcriptome sequencing, this metric can be evaluated by comparing the fraction of reads that hit each transcript in multiple sequencing runs and measuring the coefficient of variation (CV) of the read counts.
