In Silico Analyses of the Role of Codon Usage at the Hemagglutinin Cleavage Site in Highly Pathogenic Avian Influenza Genesis

Figures

• 

  Figure 1. Phylogenetic tree of all 20,448 full-length LPAIV HA nucleotide sequences. The tree was generated after HA subtype reassignment of incorrectly annotated sequences. Branch lengths are arbitrary, and clades are colored by HA subtype using the colors indicated on the right. The clade containing H17 and H18 bat sequences was collapsed and used as root. The inner to outer rings show, respectively: the HA subtype, the number of basic amino acids in the P4 to P1 region (mono-, di-, or tribasic), the region of isolation (Americas or Africa–Eurasia–Oceania, AEO), and the species in which the corresponding virus was isolated (terrestrial poultry, wild aquatic birds, other, and not determined) using the colors indicated on the right.

• 

  Figure 2. Sequence logos of nucleotide and amino acid composition of HA cleavage site regions (P6 to P1′, P1 position corresponding to the C-terminus of HA1) of different subtypes. For each HA subtype, the number of sequences used to generate the logo is indicated. Letter sizes are directly proportional to the percentage of sequences presenting a given character, as indicated on the y-axes.

• 

  Figure 3. Examples of species- or region-specific codon usage observed in H5, H7, and H10 sequences. (a) Sequence logos of sequences from H5 viruses isolated from wild aquatic birds (wild birds), terrestrial poultry (poultry), or all species combined (including 94 sequences from other or not determined species), shown for nucleotide (left) and amino acid (right) sequences. The number of sequences used to generate the logos is indicated in brackets. (b) Sequence logos of sequences from H7 viruses isolated in the Americas, the AEO region, or both combined, using the same representation as in (a). (c) Sequence logos of sequences from H10 viruses isolated from wild aquatic birds (wild birds), terrestrial poultry (poultry), or all species combined (including 66 sequences from other or not determined species), using the same representation as in (a).

• 

  Figure 4. Number of adenines in the P4 to P1 HA cleavage site region for each subtype. The number of adenines in the 4 codons preceding the HA cleavage site was counted in each sequence, and the percentage of sequences with a given number of adenines was calculated independently for each subtype. A darker shade of blue indicates a higher number of adenines, and the exact percentages are indicated above each bar.

• 

  Figure 5. Length of the longest adenine stretch in the P4 to P1 HA cleavage site region for each subtype. The longest stretch of adenines in the 4 codons preceding the HA cleavage site was identified in each sequence, and the percentage of sequences with a given stretch length was calculated independently for each subtype. A darker shade of blue indicates a longer stretch, and the exact percentages are given above each bar.

• 

  Figure 6. Number of substitutions required in sequences from each subtype to obtain a tribasic cleavage site via substitutions. For each sequence, the number of substitutions required to obtain an arginine in P1 and two other basic amino acids (lysine and arginine only) in P4 to P2 was determined. Only AGG and AGA codons were allowed for arginine, as they represent virtually all arginine codons in HPAIVs in nature. The percentage of sequences requiring a given number of substitutions was calculated independently for each subtype. A darker shade of blue indicates that fewer mutations are required, and the exact percentage is given above each bar.

• 

  Figure 7. Number of substitutions required to increase adenine stretch length in the P4 to P1 region using only codons found in the dataset at each given position in each subtype. (a) Longest adenine stretch achievable with the codon usage of each subtype. (b) Number of substitutions required for different adenine stretch lengths for H3, H5, H6, H7, and H9 HAs. For each sequence, the number of substitutions required to obtain a 4- to 8-adenine stretch is shown. The absence of a graph denotes the impossibility of obtaining an adenine stretch of the indicated length using only codons observed in nature. The percentage of sequences needing a given number of substitutions was calculated independently for each subtype. A darker shade of blue indicates that fewer substitutions are required, and the exact percentage is given above each bar.