Although the occurrence of pseudogene transcription is reported in the literature, these regions may have been mistakenly assembled and/or erroneously annotated in bacterial genomes [ 24, 25].
However, it is possible that these loci represent paralogs sharing almost identical sequences that erroneously assembled into single gene models.
This can be caused by an incomplete or erroneously assembled genome, mitochondrial genes, contamination, or simply low quality ESTs.
The SNPs falsely predicted by PolyBayes represent either experimental errors in the ESTs or differences between paralogous sequences erroneously assembled into a single contig.
High confidence genes were filtered from potentially erroneously assembled transcripts based on homology to known proteins and read coverage of the transcript.
Confusion of paralogous and allelic SNPs, misdesigned primers and failed cloning are just a few of the problems that could arise from relying on erroneously assembled sequences.
Clusters may contain more than one consensus sequence for various reasons (e.g., clone has long insert, clones are splice variants, consensus sequences are erroneously assembled).
Because of the low coverage of the nuclear genome, any nuclear-copy reads erroneously assembled into a mitochondrial contig would be at much lower frequency than the mitochondrial reads and should not affect the mitochondrial consensus.
This set of PUTs with low or no similarity to other species could represent erroneously assembled PUTs, but since estimates from the P. glauca transcriptome suggest that as much as 11.6 Mbp of the transcriptome is still missing [ 22], they could also represent previously undiscovered spruce transcripts.
To eliminate erroneously assembled transcripts, we applied a stringent criteria [ 28, 29] to identify probable transcripts which should be detected by two assemblers in same sample or identified in at least two individual samples by same assembler and defined a set of 235,283 cleavage stage transcripts.
