Bioinformatics analysis of large-scale viral sequences: from construction of data sets to annotation of a phylogenetic tree

Abstract

Due to a significant decrease in the cost of DNA sequencing, the number of sequences submitted to the public databases has dramatically increased in recent years. Efficient analysis of these data sets may lead to a significant understanding of the nature of pathogens such as bacteria, viruses, parasites, etc. However, this has raised questions about the efficacy of currently available algorithms for the study of pathogen evolution and construction of phylogenetic trees. While the advanced algorithms and corresponding programs are being developed, it is crucial to optimize the available ones in order to cope with the current need. The protocol presented in this study is optimized using a number of strategies currently being proposed for handling large-scale DNA sequence data sets, and offers a highly efficacious and accurate method for computing phylogenetic trees with limited computer resources. The protocol may take up to 36 h for construction and annotation of a final tree of about 20,000 sequences.

Figure 1. An illustration outlining the procedure demonstrated in this protocol. All the tools used in this protocol are mentioned on the left side whereas the tools that can be used in future are mentioned on the right side of the figure.

Figure 2. A twenty-year perspective of non-structural (NS) gene sequence submissions to the GenBank. The green bar represents the upsurge in sequence submission during the 2009 Swine Flu pandemic.

Figure 3. An overview of the sequence alignment as seen in the CodonCode Aligner. The floating window shows the progress of the completion of the four fundamental steps: Initialization, Overlap Detection, Alignment and Data Model Update.

Figure 4. Processing of the data set for the construction of a phylogenetic tree in the raxmlGUI program. Please note two different windows, raxmlGUI 1.1 and raxmlGUI 1.1—Pythone—132 by 15. The progress of the tree construction will be displayed in the latter window.

Figure 5. An overview of the tree annotated in FigTree. The clustering pattern of different subtypes of influenza viruses is highlighted with different colors. The subtypes such as H3, H5 and H6 made a bigger cluster owing to same genetic nature. All other subtypes had shown diffused pattern within the tree. For clarification purposes, only a tree of the avian influenza NS1 gene is displayed.

Figure 6 (See previous page). A presentation of the phylogenetic tree (A) in radial view and (B) in a circle. These different display patterns are crucial to the final interpretation of results. The tree in radial view displays a clear division of the tree into two sub-trees.