Hierarchical cluster of Ebola virus Envelope Glycoprotein

Hierarchical clustering of the Envelope glycoprotein in Ebola virus for the subtypes- Cote d’Ivoire, Zaire, Sudan and Reston- appear to have a temporal and geographical significance as seen in Figure 1. The strains of each subtype, found in similar locations, of the Ebola virus were closely clustered together. Within these clusters, more exclusive connections were observed between strains isolated from the same or within a few years of each other. The hierarchical cluster of the Envelope glycoprotein clearly exhibited a major trend towards clustering according to time and location of isolation.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 1. Hierarchical Cluster of Ebola virus Envelope glycoprotein (GP) Amino Acid Sequences. Sequences for the Envelope glycoprotein of twelve strains of Ebola virus were clustered according to their amino acid frequencies. The dendrogram displays results from the clustering analysis. Subtypes, along with the specific strain and the year of isolation, of the Ebola virus are shown in each box. Hierarchical clustering was performed with the Gene Cluster 3.0 program (Eisen) using an uncentered correlation metric and an average linkage clustering method.

 

 

 

Hierarchical cluster of Influenza virus A Hemagglutinin

A similar temporal and geographical pattern of clustering was observed in the analysis of strains of Influenza virus A (strain A, H1N1). We found that strains from host countries that are geographically related generally cluster more exclusively together than with strains from countries that are geographically distant (Figure 2). Strains can be generally classified into two categories based on geographical location: North American strains and Eastern Hemisphere strains. With few exceptions (Malaysia: Malaya-54, USA: Texas-07, Brazil-78, Chile-83, USA: Memphis-96, Australia-80, and Netherlands-85), the analysis showed that North American strains cluster together at lower levels more often than with Eastern Hemisphere strains. The same pattern of geographical clustering is seen with Eastern Hemisphere strains. A time-dependent clustering pattern is also implied from the analysis. Our dendrogram shows that as one moves from less exclusive (higher level) to more exclusive (lower level) clusters of the dendrogram, the years of isolation of included strains become closer in time.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 2. Hierarchical Cluster of Influenza virus A, subtype H1N1, Hemagglutinin Amino Acid Sequences. The protein sequences for Hemagglutinin protein from 32 isolated samples of Influenza virus A (subtype H1N1) were hierarchically clustered according to their amino acid frequencies. The dendrogram shows the results of the clustering analysis. Each entry within the dendrogram represents an independently isolated sample of strain A, H1N1. Brackets on the far right indicate the top, middle, and bottom large clusters. The samples are distinguished by their location and year of isolation. Hierarchical clustering was performed with Gene Cluster 3.0 (Eisen) using an uncentered correlation metric and an average linkage clustering method.*Country of isolation for Henry-36 could not be confirmed

 

 

 

Hierarchical cluster of Influenza virus B Glycoprotein NB

For Glycoprotein NB from the Influenza virus B (strain B), hierarchical clustering revealed a geographical and temporal trend as seen in Figure 3. The geographical trend is highly observed when following the connections of the strains from the United States (USA). While it appears that USA samples are dispersed across the dendrogram, linkage of USA samples is apparent at the highest levels and connects to USA samples found at lower levels. Additionally, Eastern Hemisphere samples were clustered in close proximity to each other in the dendrogram. The relational distances within the dendrogram between samples from East Asian countries and those of the USSR remain small. The sample from Australia is found relatively close to the Asian (East Asian and USSR) samples. The temporal trends are highly exhibited at lower levels of the dendrogram. For example in Figure 3, samples from USA: Memphis-86, Leningrad-86 and Australia: Victoria-85 are closely clustered together, as well as samples from Beijing-87 and USA: Memphis-89. Because the strains in these two clusters are geographically unrelated, their close clustering demonstrates time-dependency inherent in the organization of these strains.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 3. Hierarchical Cluster of Influenza virus B Glycoprotein NB Amino Acid Sequences. Sequences of Glycoprotein NB from 11 samples of Influenza virus B were hierarchically clustered based on the amino acid frequencies of each sample. The dendrogram shows the results of the cluster analysis. Each entry in the dendrogram represents an independently isolated sample of Influenza virus B. The samples are distinguished by their location and year of isolation. Hierarchical clustering was performed with Gene Cluster 3.0 (Eisen) using an uncentered correlation metric and an average linkage clustering method.