Table 1.
Omics-driven genome annotation summary.
Figure 1.
Schematic showing the comparative omics-based genome annotation workflow employed for annotation refinement.
Transcriptomic data generated from an unbiased universal Yersinia microarray and peptide data matched to a 6-frame genome translation were layered on existing genome annotations to validate predicted protein coding sequences and identify annotation anomalies. This evidence can be used independently or combined with putative protein identifications derived from a comparative genomics approach for genome annotation refinement.
Figure 2.
Distribution of oligos for universal Yersinia microarray.
A universal array was designed to represent genes from seven different Yersinia strains on a single chip. A) A representative sample of oligos and their purported mapping to genes from each of the seven Yersinia strains represented on the array is shown. The number of strains (n = x) for which a given oligo corresponds to an annotated protein coding sequence is indicated, illustrating a subset of the possible combinations of unique or shared orthologs at the oligo level. B) The distribution of oligos based on existing annotations is provided; oligos were predicted to be either unique to any one of the seven strains (n = 1) or shared by multiple strains (n>1).
Figure 3.
Categorization of experimental evidence.
Preliminary analysis shows the distribution of peptide and oligo evidence across annotated open reading frames (ORFs) (i.e., predicted proteins) in black and unannotated ORFs in grey. Solid regions indicate ORFS with complementary oligo and peptide evidence and hashed regions show ORFs with only peptide evidence. A) shows ORFs exhibiting two or more unique non-redundant peptides and B) shows ORFs exhibiting a single unique non-redundant peptide.
Figure 4.
Representative identification of a novel gene.
In the upper frame, gene level alignments are shown for the three strains examined in this study. Predicted protein coding genes are shown in grey and oligo evidence is shown in black. Sequence alignments are shown in the lower frame. Annotated protein sequences are underlined, and experimental peptide evidence is shown in black text. For YPDSF, this region reveals substantial oligo and peptide evidence in an unannotated region indicating a novel gene.
Figure 5.
Representative identification of an incorrectly predicted translational start site.
Protein sequence alignments are shown for the three strains examined (YPO, YPDSF, and YPTS) in addition to several other Yersinia species. Annotated protein sequences are underlined and experimental peptide and oligo evidence is shown in black text. For YPO0453, evidence flanks the predicted start site (shown by boxed region) and the observed peptide sequence reveals translation of a leucine instead of methionine confirming an N-terminal extension.
Figure 6.
Representative identification of an expressed pseudogene.
Pseudogenes are considered translationally silent and typically excluded from protein databases. YPO1195 is categorized as a pseudogene due to the presence of a stop site mid sequence (shown by vertical black bar). Both oligo evidence (black boxes, upper frame) and peptide evidence (black text, lower frame) were observed on either side of the predicted stop codon, indicating expression of this feature.
Figure 7.
Representative identification of a putative translational frameshift.
A) Hybridization evidence for oligos labeled A, B, and C is shown. Expression levels are shown normalized to each oligo's mean (via a Z-score calculation) across a time course/thermal switch (37°C/26°C) experiment for Y. pestis CO92 (YPO) and Y. pestis Pestoides F (YPDSF). Green indicates down-regulation relative to the mean, and red indicates up-regulation relative to the mean. Genome annotations are labeled for YPDSF_1005 and YPO2124 corresponding to annotated coding sequences. Although oligos A and B purportedly lack a corresponding transcription for YPDSF (NA = not applicable), evidence clearly shows hybridization consistent with oligo C. B) illustrates the 210 aa translation of YPO2124 and C) illustrates the 63 aa translation of YPDSF_1005. Frame translations are shown below gene level detail with oligo evidence (black) overlaid on each gene and peptide evidence (red) overlaid on the appropriate reading frame. For YPDSF_1005, gene alignment with YPO2124 reveals two oligos upstream of the coding region. Corroborating peptide evidence was also seen upstream but in a different reading frame than the existing annotation. This evidence supports the expression of YPDSF_1005 as a frameshifted protein.
Figure 8.
Venn diagram overlap of evidence for unannotated proteins identified by single peptide identifications.
Open reading frames with evidence were initially filtered based on the presence of a non-redundant peptide. All singlet peptides were required to have corroborating oligo evidence or it was required that the singlet peptide mapped to an orthologous peptide in one of the alternate strains examined.