Section B. Comparative Genomics Analysis of 2013 H7N9 Influenza A Viruses. Objective

Transcription

1 Section B. Comparative Genomics Analysis of 2013 H7N9 Influenza A Viruses Objective Upon completion of this exercise, you will be able to use the Influenza Research Database (IRD; to: Search for virus sequences and view detailed information about these sequences in IRD Save selected sequences as a working set in your private Workbench space Combine multiple working sets Build a phylogenetic tree on a set of sequences to infer their evolutionary relationships Use Meta-CATS to identify nucleotide or amino acid positions that significantly differ between groups of virus sequences Perform a multiple sequence alignment to observe sequence conservation and variations Determine if significant positions are located in viral protein Sequence Features and examine Sequence Feature Variant Type reports Search for 3D protein structures and highlight Sequence Features and custom positions on a structure Background H7 viruses normally circulate in birds and horses. Before March 2013, a search for H7 influenza strains in IRD returned a total of 1485 strains in IRD, with 1306 from birds, 102 from environmental samples (usually bird droppings), 33 from horses, and only 15 from humans (11 H7N7, H7N1, H7N2, 2 H7N3). No human isolates were H7N9. In March 2013, several cases of Influenza virus A H7N9 subtype were identified in Shanghai, China and surrounding provinces. From February to May, 2013, a total of 133 cases have been confirmed, including 37 deaths. Since October 2013, a second wave of human cases has been occurring, a total of 74 as of January 21, Fortunately, no evidence of ongoing human-tohuman transmission for the current H7N9 outbreak has yet been found. This suggests that the virus is undergoing stuttering transmission in which a virus that normally circulates in an animal reservoir infects a person, but further human-to-human transmission does not occur. In general, viruses capable of stuttering transmission have acquired novel sequence variations that allow them to infect humans (human adaptation), but have yet to acquire sequence variations that allow them to sustain efficient transmission between humans. We will perform an in-depth statistical analysis using sequence records and analysis tools available in IRD ( to clarify the origin of the HA segment and to identify candidate sequence variations that might be involved in this type of human adaptation. 4

2 Analysis Workflow Search for sequences and save sequences into working sets: - search for H7N9 HA nucleotide sequences and save sequences as working set (1) - BLAST for HA nucleotide sequences similar to H7N outbreak sequences and save them as working set (2) - combine working sets (1)-(2) into one (3) - convert nucleotide working set (3) into protein working set (4) Construct nucleotide phylogenetic tree: - construct phylogenetic tree using working set (3) - color tree to reveal host and subtype specific branching patterns Run Metadata-driven Comparative Analysis Tool (Meta-CATS): - input the protein working set (4) to Meta-CATS - group the sequences into two groups: human H7N outbreak isolates, and similar older H7 isolates - identify positions that are significantly different between the two sets - convert position coordinates on the H7 numbering scheme into coordinates on the H3 numbering scheme Visualize multiple sequence alignment: - align HA protein sequences - observe the variant positions on the alignment Determine if the significant positions are located in Sequence Features: - follow the Sequence Feature linkage on the Meta-CATS report - examine Sequence Features containing the significant positions Highlight significant positions and Sequence Features on protein structure: - search for H7 HA 3D protein structures - highlight Meta-CATS positions and Sequence Features on a structure 5

3 I. Search for sequences and save matching sequences into working sets 1. Search for H7N9 HA sequences using structured search interfaces a. Go to the IRD homepage ( mouse-over Search Data in the grey navigation bar, then Search Sequences and click Nucleotide Sequences. b. The Nucleotide Sequence Search page allows you to search for sequences based on data type, virus type, subtype, strain name, segment, host, geographical region, complete sequences or not, H1N1 pandemic sequences or not, and date range. c. For this exercise, we are going to search for HA segment sequences from H7N9 strains. Select the following criteria and click the orange Search button to run the query. Virus Type: ý A Complete Sequences: ý Complete Sequences Only Select Segments: ý 4 HA Advanced Options: ý Remove Duplicate Sequences Sub Type: H7N9 d. Note that IRD shows instant count of search results here to help you search quickly and efficiently. When you select search criteria on search pages, you will instantly know how many records match your search criteria without clicking the Search button and actually running the search. e. The Search Results page will be displayed. Here you can: i. Save the search query to your Workbench and rerun the search again later. 6

4 ii. Download the sequences (gene, CDS, protein) by clicking Download. iii. Select records and run an analysis on the selected records by mousing-over the Run Analysis button and clicking a desired analysis option. iv. Store selected sequences as a working set in the Workbench so that you can run various analyses on the working set. v. View the details for any item in the results table by clicking on View next to any row. f. Now click the Collection Date header in the results table to sort records by date. If you need advanced sorting options and want to display additional fields, click the Display Settings About Us Community Announcements Links Resources Support Workbench Sign In button. SEARCH DATA ANALYZE & VISUALIZE WORKBENCH SUBMIT DATA Home Nucleotide Sequence Search Results Nucleotide Sequence Search Results Your Selected Items: 95 items selected Deselect All Add to Working Set Save Search Run Analysis Download Your search returned 95 segments. Search Criteria Displaying 50 records per page, sorted by Strain Name in Select all 95 segments ascending order. 4 HA CY Yes 1697 H7N Blue-Winged How many HA segment sequences did you find from the current outbreak? 4 HA KF Yes 1710 H7N9 04/29/2013 Human China -N/A- -N/A- *A/Changsha/2/2013(H7N9) g. To analyze these sequences, we will select records by ticking the checkbox and adding them to a 4 HA CY * Yes 1683 H7N9 05/03/2013 Chicken/Avian China -N/A- -N/A- *A/chicken/Guangdong/SD641/2013(H7N9) working set by clicking 4 HA the KF Add * No to 1696 working H7N9 04/2013 set button. Chicken/Avian This China way, -N/A- we -N/A- will be A/chicken/Jiangsu/1021/2013 able to retrieve the 4 HA CY Yes 1683 H7N9 04/16/2013 Chicken/Avian China -N/A- -N/A- A/chicken/Jiangsu/S002/2013 data from the Workbench later and run various analyses on the same data set. h. You ll be prompted 4 to HAlog CY in to Yes your 1683 Workbench H7N9 04/16/2013account Chicken/Avian in China order -N/A- to save -N/A- data A/chicken/Jiangsu/SC537/2013 to a working set. If 4 HA CY Yes 1683 H7N9 05/03/2013 Chicken/Avian China -N/A- -N/A- A/chicken/Jiangxi/SD001/2013 you don t have an account already, simply register for an account for free by choosing the 4 HA KF * No 1683 H7N Chicken/Avian China -N/A- -N/A- *A/chicken/Rizhao/515/2013(H7N9) 4 HA KF * No 1683 H7N Chicken/Avian China -N/A- -N/A- A/chicken/Rizhao/713/2013 Register for a new account option and following the prompts. 4 HA KF Yes 1683 H7N9 04/2013 Chicken/Avian China -N/A- -N/A- A/chicken/Shanghai/017/2013 i. A lightbox of Add 4 to HA Working CY Yes Set 1683 will H7N9 pop 04/03/2013 up. Chicken/Avian Now create China a -N/A- new -N/A- working A/chicken/Shanghai/S1053/2013 set and name it 4 HA CY Yes 1683 H7N9 04/03/2013 Chicken/Avian China -N/A- -N/A- A/chicken/Shanghai/S1055/2013 H7N9 HA complete sequences. Click Add to Working Set to save the sequences to a 4 HA CY No 1683 H7N9 04/03/2013 Chicken/Avian China -N/A- -N/A- A/chicken/Shanghai/S1076/2013 working set. 4 HA CY Yes 1683 H7N9 04/03/2013 Chicken/Avian China -N/A- -N/A- A/chicken/Shanghai/S1077/ BLAST for HA sequences similar to H7N outbreak sequences Display Settings 1 2 Next > Page: 1 of 2 More columns were returned than can be displayed without scrolling. Use scroll bars at top and bottom of display to move right and left or reduce the number of columns displayed by using the Display Settings link above. Segment Protein Name Sequence Accession Complete Segment Genome Length Subtype * Collection Date 4 HA HQ No 1698 H7N9 01/26/2008 *Eurasian Teal/Avian 4 HA FN No 1683 H7N9 01/26/2008 *Eurasian Teal/Avian Host Species Country State/Province Now we are going to expand 4 HA KF the sequence * No 1683set H7N9 by including 2013 Chicken/Avian HA sequences -N/A- -N/A- that -N/A- are A/chicken/Zhejiang/PA-DTID-ZJU01/2013 highly similar to the 4 HA CY Yes 1683 H7N9 04/22/2013 Chicken/Avian China -N/A- A/chicken/Zhejiang/SD007/2013 outbreak sequences. We 4 will HA CY select a Yesrepresentative 1683 H7N9 04/11/2013 isolate Chicken/Avian and perform China -N/A-a BLAST -N/A- A/chicken/Zhejiang/SD033/2013 search of nucleotide 4 HA CY Yes 1683 H7N9 04/16/2013 Duck/Avian China -N/A- -N/A- A/duck/Anhui/SC702/2013 sequences. The IRD BLAST tool utilizes the NCBI BLAST program set and has a collection of custom influenza sequence databases to search against. a. Select A/Shanghai/02/2013 from the results table, mouse over Run Analysis and click BLAST. 7 Flu Season (SOP) Strain Name Spain -N/A *A/Anas crecca/spain/1460/2008(h7n9) Spain -N/A *A/Anas crecca/spain/1460/2008(h7n9) 4 HA KF No 1683 H7N null China -N/A- -N/A- A/Anhui/PA-1/ HA CY Yes 1731 H7N9 02/07/2008 *Blue-Winged Teal/Avian 4 HA CY Yes 1731 H7N9 03/05/2008 *Blue-Winged Teal/Avian Teal/Avian Guatemala -N/A- -N/A- A/blue-winged teal/guatemala/cip049-01/2008 Guatemala -N/A- -N/A- A/blue-winged teal/guatemala/cip049-02/2008 USA Ohio -N/A- *A/blue-winged teal/ohio/566/2006(h7n9) 4 HA KF Yes 1710 H7N9 03/22/2013 Human China -N/A- -N/A- *A/Changsha/1/2013(H7N9) 4 HA CY Yes 1683 H7N9 04/16/2013 Chicken/Avian China -N/A- -N/A- A/chicken/Jiangsu/SC035/ HA CY Yes 1683 H7N9 04/16/2013 Chicken/Avian China -N/A- -N/A- A/chicken/Jiangsu/SC099/ HA KF * No 1683 H7N Chicken/Avian China -N/A- -N/A- A/chicken/Rizhao/719b/ HA CY Yes 1683 H7N9 04/03/2013 Chicken/Avian China -N/A- -N/A- A/chicken/Shanghai/S1358/ HA CY Yes 1683 H7N9 04/03/2013 Chicken/Avian China -N/A- -N/A- A/chicken/Shanghai/S1410/ HA CY Yes 1683 H7N9 04/03/2013 Chicken/Avian China -N/A- -N/A- A/chicken/Shanghai/S1413/ HA KC No 1683 H7N9 04/2013 Chicken/Avian China -N/A- -N/A- A/chicken/Zhejiang/DTID-ZJU01/2013

5 b. In the Select Sequence Type lightbox, select Nucleic Acid (Segment) and click Continue. c. Now the BLAST setting page is loaded. IRD provides a collection of custom influenza sequence databases to search against using BLAST. Select Blastn, Nucleotides for segment 4 HA, and 100 results to display. For the other parameters keep the default settings. Click Run. SEARCH DATA ANALYZE & VISUALIZE WORKBENCH SUBMIT DATA Home Nucleotide Sequence Search Results Identify Similar Sequences (BLAST) Identify Similar Sequences (BLAST) Compare sequences you provide or select from the IRD database consisting of selected sets of sequences from the IRD database or create your own database for comparison from a working set on your workbench. Note: An asterisk (*) = required field ANALYSIS NAME FORMAT OF SEQUENCES PROVIDED Blastn (nucleotide) Blastx (search protein using nucleotide translated in 6 reading frames) SELECT DATABASE TO SEARCH * Use generated database Nucleotides for segment 4 H1 Nucleotides for segment 4 H3 Nucleotides for segment 4 H5 Nucleotides for segment 4 HA Nucleotides for segment 5 NP Nucleotides for segment 6 N1 Tip: To select multiple or deselect, Ctrl-click (Windows) or Cmd-click (MacOS) INPUT SEQUENCES 1 record was previously selected from search results Use working set to run blast OUTPUT FORMAT Remove data from genetically manipulated strains from the result Number of Results to Display for each Input Sequence d. On the BLAST Report page, all nearest hits are listed in the table. Click a hit to view its alignment. Click the IRD link (e.g., ird ) to view the hit s segment/ protein details page in IRD. What About Us Community Announcements Links Resources Support Workbench Sign In are the host and subtype of the sequences that are most similar to the H7N9 outbreak sequences? SEARCH DATA ANALYZE & VISUALIZE WORKBENCH SUBMIT DATA Home Nucleotide... Results Identify Similar Sequences (BLAST) Results BLAST Report BLASTN [Sep ] Reference: Altschul, Stephen F., Thomas L. Madden, Alejandro A. Schaffer, Jinghui Zhang, Zheng Zhang, Webb Miller, and David J. Lipman (1997), "Gapped BLAST and PSI-BLAST: a new generation of protein database search programs", Nucleic Acids Res. 25: Query: gb:kf Organism:Influenza#A#virus#A/Shanghai/02/2013 Segme Database: Influenza nucleotide sequences Seg4 70,570 sequences; 95,790,805 total letters Save Analysis Id Add to Working Set Sequence header >ird Country:China Influenza A virus (A/Shanghai/02/2013(H7N9)) segment 4 hemagglutinin (HA) gene, complete cds. gb KF >ird Country:China Influenza A virus (A/Zhejiang/DTID-ZJU01/2013(H7N9)) segment 4 hemagglutinin (HA) gene, complete cds. gb KC Bit Score E Value >ird Country:China Influenza A virus (A/Fujian/1/2013(H7N9)) segment 4 hemagglutinin (HA) gene, complete cds. gb KC >ird Country:China Influenza A virus (A/environment/Hangzhou/34/2013(H7N9)) segment 4 hemagglutinin (HA) gene, complete cds. gb KF e. Return to the BLAST >ird Report Country:China page Influenza by A virus (A/Hangzhou/2/2013(H7N9)) clicking the segment Results 4 hemagglutinin breadcrumb. (HA) gene, complete Now 3305 select 0.0 all hits by cds. gb KF ticking the checkbox >ird Country:China in the Influenza column A virus (A/Nanjing/1/2013(H7N9)) header and segment 4 click hemagglutinin Add (HA) gene, complete to cds. gb KC Working 3305Set 0.0 to add these >ird Country:China Influenza A virus (A/environment/Nanjing/2913/2013(H7N9)) segment 4 hemagglutinin (HA) gene, complete cds. gb KC sequences to a new working set named A/Shanghai/02/2013 BLAST hits. 3. Construct segment and protein working sets a. Now we are going to combine the working sets of H7N9 HA complete sequences and >ird Country:Taiwan Influenza A virus (A/Taiwan/T02081/2013(H7N9)) segment 4 hemagglutinin (HA) gene, partial cds. gb KF A/Shanghai/02/2013 BLAST hits and use the combined working set for downstream analyses >ird Country:China Influenza A virus (A/Hangzhou/1/2013(H7N9)) segment 4 hemagglutinin (HA) gene, complete cds. gb KC >ird Country:China Influenza A virus (A/Shanghai/4664T/2013(H7N9)) segment 4 hemagglutinin (HA) gene, complete cds. gb KC >ird Country:China Influenza A virus (A/Zhejiang/HZ1/2013(H7N9)) segment 4 hemagglutinin (HA) gene, complete cds. gb KF >ird Country:China Influenza A virus (A/Hangzhou/3/2013(H7N9)) segment 4 hemagglutinin (HA) gene, complete cds. gb KF >ird Country:Taiwan Influenza A virus (A/Taiwan/S02076/2013(H7N9)) segment 4 hemagglutinin (HA) gene, partial cds. gb KF >ird Country:China Influenza A virus (A/chicken/Zhejiang/DTID-ZJU01/2013(H7N9)) segment 4 hemagglutinin (HA) gene, complete cds. gb KC >ird Country:China Influenza A virus (A/duck/Zhejiang/12/2011(H7N3)) segment 4 hemagglutinin (HA) gene, complete cds. gb JQ >ird Country:China Influenza A virus (A/duck/Zhejiang/11/2011(H7N3)) segment 4 hemagglutinin (HA) gene, complete cds. gb JQ >ird Country:China Influenza A virus (A/duck/Zhejiang/2/2011(H7N3)) segment 4 hemagglutinin (HA) gene, complete cds. gb JQ >ird Country:China Influenza A virus (A/duck/Zhejiang/10/2011(H7N3)) segment 4 hemagglutinin (HA) gene, complete cds. gb JQ >ird Country:China Influenza A virus (A/duck/Zhejiang/DK16/2013(H7N3)) segment 4 hemagglutinin (HA) gene, complete cds. gb KC

6 Click the Workbench tab from the grey navigation bar to go to your Workbench. You ll see the saved working set listed at the top of the Workbench table. b. Click the checkboxes for the two working sets we just saved. Click More Actions then Combine. Name the combined working set to H7N9 HA complete seqs + blastn hits. This working set contains the HA segment sequences from H7N9 isolates and sequences that are most similar to the HA segment sequences of the current H7N9 outbreak. You will notice the duplicate sequences are automatically removed. c. Next, we are going to convert the combined segment working set into a protein sequence working set. To do so, select the combined working set. Click More Actions then Convert. d. In the Convert Working Set lightbox, select Protein as data type and name the working set to: H7N9 HA complete seqs + blastn hits protein. Click Convert. e. Access your Workbench by clicking the Workbench tab. You will see the newly created working sets at the top of the content list. 4. (Optional) Combine custom sequences with IRD sequences When you have custom sequences and want to analyze your sequences along with IRD sequences, you can upload your sequence file to your Workbench and then combine the uploaded file with a saved working set. a. Prepare your sequences in FASTA format. b. Go to the Workbench, click More Actions then Upload File. c. Next, name your file, choose File Type (i.e., Unaligned FASTA ) and Data Type, click Browse and select the desired FASTA file on your computer. Choose the desired Fasta File Definition Line format. Click Upload. The uploaded file will be displayed in the Workbench table. d. Now select desired working set. Click More Actions then Combine with Uploaded File. e. In the pop-up lightbox, choose the desired sequence file from the list of uploaded files, name the combined working set, and click Combine. You will see the newly combined working sets at the top of the content list. II. Construct an HA segment phylogenetic tree a. Now we will construct a phylogenetic tree using HA segment sequences from H7N9 isolates and segment sequences that are most similar to the current H7N9 outbreak isolates. b. We are going to add HA sequence from A/ruddy turnstone/de/2378/1988 as an outlier to the H7N9 HA and closely related sequences. 9

7 i. Mouse over the Search Data tab and click Nucleotide Sequences. ii. iii. On the Nucleotide Sequence Search page, select 4 HA from the Select Segments list, type A/ruddy turnstone/de/2378/1988 in the strain name box, and click Search to run the search. The search result page will show 1 segment. Select the segment by ticking the checkbox next to it. Then click the Add to working set button and add it to the working set: H7N9 HA complete seqs + blastn hits. c. On the Workbench page, click View next to H7N9 HA complete seqs + blastn hits. d. The working set details page displays the sequence records saved in the working set. Select all records by clicking the checkbox above the table. Mouse over Run Analysis and click Generate Phylogenetic Tree. e. On the Tree setting page, select Quick Tree, choose strain name and date as tree tip label, and click Build Tree. Generate Phylogenetic Tree Tutorial The "Quick Tree" option uses PhyML [ Guindon, S. and Gascuel, O., (2003) Syst Biol. 52: ] and IRD-defined settings to infer phylogenies based on sequences for datasets of at most 1000 sequences. The "Custom Tree" option offers a choice between the PhyML or RaxML [Stamatakis, A. et al. (2005) Bioinformatics 21: ] algorithms, and the ability to define parameter settings. The Custom Tree option, using RaxML, must be used for datasets exceeding 1000 sequences. Click here to view a tutorial on generating a phylogenetic tree using IRD tools. ANALYSIS NAME H7N HA segment phylogeny TREE GENERATION Quick Tree (Let IRD set all parameters - view all parameters) Custom Tree (I want to set my own parameters and/or I have a large dataset) INPUT 146 SEGMENTS SELECTED FOR TREE Exclude 1 segments with poorly pre-aligned sequences Use all 146 segments and run MUSCLE to align them LABEL TREE TIPS (ENDS) WITH Strain Name Specify custom format of tip label (max 4) Strain Name Accession Number Date Country USA State Segment Protein Symbol Season Type SubType Host Species 2009 ph1n1-like Phenotype Markers Clear Build Tree f. While the analysis is running, you can save the analysis to your Workbench by entering a name and then clicking Save to Workbench. Once it is saved, you can come back to the Workbench at any time to retrieve the analysis results. g. After the analysis is finished, a View Phylogenetic Tree page will be loaded. Here you can save the phylogenetic file in Newick or PhyloXML format to your computer. Click View Tree to load the Archaeopteryx Tree Viewer window. 10

8 View Phylogenetic Tree Save Analysis Newick File PhyloXml File Phylip File Tree Parameters PhyML Log Tree Build Parameters Click the "View Tree" button below to launch the tree viewer software in a new window. If you prefer other viewing software, the tree data is available for download in Newick or PhyloXml format using the buttons above. Due to security concerns, certain browsers (e.g. Safari and Firefox) have disabled Java plug-ins by default. If the Tree Viewer takes a long time to load, please test your browser's Java plug-in to make sure it can display Java Applets properly. Safari has recently tightened the security settings on Java Applets, which may affect image export functions of the Tree Viewer. Click Here for instructions on how to fix this. ENHANCED TREE VIEWER The IRD team provides software that allows 'decoration' of your tree by features such as host species, year, country, and subtype. This custom software is based on Archaeopteryx. In the tree viewer, use the drop-down menu for basic decoration or advanced decoration to select the feature for coloring. The decorated tree and corresponding legend can be exported using options in the File drop-down menu. A user's guide is available. How to create a publication quality tree image View Tree h. A Tree Viewer window will pop up. Many tree customization options exist including: reroot the tree, collapse/expand/display subtree, swap descendants, decorate (color) the tree leaves by any associated metadata (e.g. host or year of isolation, etc.), resize the tree, zoom in/out, fit the tree to window, change the font size, etc. i. In the Tree Decorations section, select HA&NA Subtype from the Basic Decoration Options. Click Show Legend to display the color code for different subtypes. ii. The default colors may or may not be ideal for your purpose. You can change the color by using the Advanced Decoration. In the Advanced Decoration Options dialog box, select HA&NA Subtype, click the Manual Decoration checkbox and click Go. iii. iv. Check H7N9 and choose red in the color palette, then click Apply. Now the H7N9 strains are colored in red. The tree shows that the HA sequences from the H7N9 outbreak are similar to the HA sequences from a series of duck H7N3 isolates from Zhejiang and a series of poultry H7N2/3/7 isolates mostly from Wenzhou, suggesting that at least two independent interspecies transmissions occurred and the HA segment of the new H7N9 isolates was likely derived from an H7N3 ancestor. v. You can save the tree image by clicking the File menu and then a file format. i. Return to the Tree Results page. Save the tree analysis to your Workbench by clicking Save Analysis. Rename the analysis so that you can recognize it later, for example, H7N9 HA segment phylogeny. Then click Save. 11

9 j. Go to your Workbench. You can see the tree is listed at the top of the Workbench table. Click View to retrieve the tree analysis result. The parameters used to generate the tree are also saved. III. Metadata-driven Comparative Analysis Tool for Sequences (Meta-CATS) Metadata-driven Comparative Analysis Tool for Sequences (Meta-CATS) A unique comparative genomics analysis tool in IRD to identify nucleotide /amino acid positions that significantly differ between two or more groups of virus sequences. Meta-CATS consists of three parts: a multiple sequence alignment (using MUSCLE), a chi-square goodness of fit test to identify positions (columns) of the multiple sequence alignment that significantly differ from the expected (random) distribution of residues between all metadata groups, and a Pearson's chi-square test to identify the specific pairs of metadata groups that contribute to the observed statistical difference. Picket BE, et al. (2013) "Metadata-driven Comparative Analysis Tool for Sequences (meta- CATS): an Automated Process for Identifying Significant Sequence Variations Dependent on Differences in Viral Metadata." Virology, 447(1-2): doi: /j.virol

10 Now we will use Meta-CATS to identify amino acid positions on the HA protein that significantly differ between the human isolates from the current outbreak and older H7 isolates. a. We are going to analyze the protein sequences we saved previously in the working set: H7N9 HA complete seqs + H7N9 blastn hits protein. So go to your Workbench, find the working set and click View to display the sequences. b. Sort the list by the Date column. Now select the following protein sequence records: i. All human H7N9 from 2013 ii. All non-human H7 isolated before 2012 iii. Mouse over Run Analysis and click Metadata-driven Comparative Analysis Tool. c. You are taken to the meta-cats tool setting page. Here we will separate these sequences into two groups according to the phylogenetic tree analysis: 2013 H7N9 isolates as a group and the rest sequences as another group. We can do so by grouping the sequences by year. Select Auto Grouping, and then select Year from the dropdown list. Now enter year break point 2012 to get groups of: (1) 2012 & before (Eurasian H7 ancestral isolates), and (2) > 2012 (human H7N human outbreak isolates). Select Unaligned FASTA, use 0.05 as our significance cutoff value and click Continue. d. On the next page, you will see the sequences are separated into two groups: Group 1 containing <=2012 sequences, and Group 2 containing >2012 sequences. Click the Run button. SEARCH DATA ANALYZE & VISUALIZE WORKBENCH SUBMIT DATA Home My Workbench Working Set (H7N HA nr + BLAST protein) Metadata-driven Comparative Analysis Tool for Sequences (meta-cats) Metadata-driven Comparative Analysis Tool (meta-cats) - Setup Subset Mouse click on the sequences to select them, Ctrl and/or shift key enable the multiple selection. After selecting the sequences, click on "Add" to move into the group. "Remove" will send the selected sequences back to the main list. Auto grouping on: Year Group1: <= Group2: > MAIN LIST OF SEQUENCES LIST OF SEQUENCES GROUP 1 Add Remove -N/A- -N/A- Q3KRH9 A/mallard/Sweden/91/02 AY N/A- -N/A- Q0A3N8 A/turkey/Minnesota/1/1988 CY N/A- -N/A- A7IRT3 A/blue-winged teal/ohio/566/2006 CY N/A- -N/A- B9X1H5 A/duck/Mongolia/119/2008 AB N/A- -N/A- C7C2V7 A/Anas crecca/spain/1460/2008 FN HA -N/A- E2JIX6 A/goose/Czech Republic/1848-T14/2009(H7N9) HQ HA -N/A- D0FI67 A/goose/Czech Republic/1848-K9/2009(H7N9) GU LIST OF SEQUENCES GROUP 2 Add Remove HA -N/A- M9TFV7 A/Zhejiang/DTID-ZJU01/2013 KC HA -N/A- N0D493 A/Fujian/1/2013 KC HA -N/A- M4YV75 A/Hangzhou/1/2013 KC HA -N/A- R4JCY8 A/Hangzhou/3/2013 KF HA -N/A- R4JC40 A/Hangzhou/2/2013 KF HA -N/A- R4SCY6 A/Taiwan/S02076/2013 KF HA -N/A- R4J9D5 A/Nanjing/1/2013 KC sequences 32 sequences Clear Run e. This analysis may take a few minutes to finish. You can save the analysis to your Workbench and retrieve it later. To do so, enter in a name (Ex., human H7N HA vs. older H7) and click Save to Workbench. 13

11 f. The Meta-CATS analysis result has two reports: a Chi-square Test of Independence result table listing the positions that have a significant non-random distribution between your specified groups, and a Pearson's chi-square test result table listing the specific pairs of groups that contribute to the observed statistical difference. Since this analysis only deals with two groups of sequences, we will primarily focus on the first result table. g. Review the Chi-square test results to see the positions that differ significantly between the current H7N9 outbreak isolates and other isolates. The residue diversity column lists the counts for each residue within a group. Now sort the results by the Chi-square value to push the most different positions to the top of the table. What is the position number with the highest Chi-square value? Position Chi-square Value P-value Degree Freedom Residue Diversity Sequence Feature 235* E-15 2 group1(33 Q) group2(1 I, 31 L) 195* E-15 2 group1(1 A, 32 G) group2(32 V) E-15 2 group1(17 D, 16 K) group2(32 S) E-15 1 group1(33 A) group2(32 V) E-15 1 group1(33 N) group2(32 D) 410* E-13 2 group1(2 N, 14 S, 17 T) group2(32 N) 321* E-10 4 group1(12 E, 1 K, 1 P, 4 R, 15 T) group2(32 R) 198* E-11 2 group1(5 A, 2 N, 26 T) group2(32 A) View SF View SF N/A N/A N/A N/A N/A View SF h. Some publications record H7 positions on an H3 numbering scheme. To compare the variant positions identified by meta-cats with positions reported in publications, we will convert the position numbers of the current analysis to the H3 numbering scheme by selecting the H3 A/Aichi/2/1968(H3N2) reference strain from the Reference Coordinate dropdown menu. i. The numbering scheme from the custom alignment of the input sequences are now mapped to both the H3 HA0 (uncleaved) and H3 HA1/HA2 (cleaved) coordinates. j. Save the analysis result to your Workbench by clicking the Save Analysis button. 14

12 IV. Visualize protein sequence alignment Now we are going to view the protein sequence alignment to confirm the meta-cats results and to verify clade relationships inferred from the phylogenetic analysis. a. From the meta-cats report page, click Visualize Aligned Sequences at the top of the page. Note: You can also run an alignment on the saved sequence working set by navigating to the working set in your Workbench area and then clicking the Visualize Aligned Sequences option from the Run Analysis pull down menu. b. The alignment is presented in the JalView visualization window. The window is interactive. i. The consensus sequence is shown at the bottom of the window. You can choose to show sequence logos by right-clicking on consensus and then selecting "Show logo". ii. iii. You can manually adjust the alignment and display using various gray menu options. Scroll right up to the region of Several amino acid substitutions, including G195V and Q235L/I are observed in all recent H7N9 isolates, but are absent from the older H7 proteins. iv. We can change the View Option to Conserved vs. reference such that only the first sequence shows full characters, for the remaining sequences, only the nucleotides/residues differing from the reference sequence are shown as full characters. v. You can download the input sequences or alignment in various formats, or save the alignment to your Workbench. Click Save Analysis, give it a name, and click Save. V. Determine if the significant positions are located in Sequence Features Sequence Features (SFs) are defined as interesting protein regions with known structural or functional properties. They are obtained from literature and other databases and validated by domain experts. Once a Sequence Feature region has been defined, the number of distinct amino acid sequences 15

13 observed in the sequence database are determined and each defined as a unique variant type. The reference strain is always Variant Type 1. The Sequence Feature (SF) column in the meta-cats table provides a convenient linkout to a list of all Sequence Features that contain that amino acid position. a. Return to the meta-cats report page. Click the View SF link for residue position 195. How many Sequence Features are mapped to this position? Custom Amino Acid Position Amino Acid Position in SF Reference Strain Your Selected Items: 0 items selected Download Save Search Your search returned 1 record. Search Criteria Displaying 50 records per page, sorted by Feature Identifier in Display Settings ascending order. Select the 1 results Feature Identifier Sequence Feature Name Sequence Feature Length Variant Types Category Amino Acid Position Evidence Comments Influenza A_H7_SF4 Influenza A_H7_determinants-ofreceptor-binding_194(5) 5 35 functional 194(178 HA1)-198 PubMed: Atypical European viruses show G186V whereas the N. American strains show G186A or G186E. b. Click View for SF4 to get to the Sequence Feature (SF) Details page. i. This SF is determinants of receptor binding. It begins at residue 194 and is 5 amino acids long. What is the reference strain used to define the position coordinates of this SF? What is the position range on the reference strain? ii. How many Variant Types does this SF have? iii. All 2013 outbreak isolates have a Valine at position 195 and an Alanine at position 198. Now we are going to search for all strains harboring V195 and A198. Click Find a VT, fill all positions with wildcards and fill in V at 195 and A at 198. Then click Search. iv. Did you find a VT harboring SVSTA in the loop? Click strain count for the VT. Are they from the current outbreak? SEQUENCE FEATURE DEFINITION Protein Name Sequence Feature Name Sequence Feature ID VT-1 Strain (reference strain) Reference Sequence Accession Reference Position HA Influenza A_H7_determinants-of-receptor-binding_194(5) Influenza A_H7_SF4 A/turkey/Italy/220158/2002(H7N3) AY (178 HA1)-198 SOURCE STRAIN(S) Source Strain VT Number Source Position Source Accession 3D Protein Structure Publication Evidence Codes Comment A/duck/HONG KONG/293 /1978(H7N2) VT U N/A- PubMed: EXP Atypical European viruses show G186V whereas the N. American strains show G186A or G186E. VARIANT TYPES Excel Download Find a VT(s) Phylogenetic tree view disabled because there are not enough variant types to generate the tree. Edit specific positions in this VT-1 sequence with IUPAC symbols or use "?" as a wild-card. If necessary, use the horizontal scroll bar to access the entire SF. Click Search to find VT(s) conforming to the edited sequence. Click Reset to restore this panel to the default VT-1 sequence. Search Enter Sequence Variation to Find Reset? V?? A Fill wildcards Sequence Variation Strain Count Variant Type 388 VT-1 14 VT S G S T T V A Total Variations

14 VI. Highlight variant positions and Sequence Features on protein structure IRD imports experimentally-determined virus protein structures from the Protein Data Bank, integrates data from IEDB and UniProt and provides various visualization options. To investigate the structural implications of the sequence variants identified by the IRD meta-cats analysis, we are going to highlight the positions on a related H7 protein structure. a. From the grey navigation bar, mouse over Search Data and click 3D Protein Structures. b. Search for the 3D structures of influenza A HA subtype H7. Virus Type: ý A Subtype: H7 Select Proteins to search: ý 4 HA c. The Search Results page displays a list of matching structures. We are going to examine the HA structure of H7N3 subtype, so click View Structure for 1TI8 to display the structure. d. Now we are on the 3D protein structure viewer page. Click and drag with your mouse in display window to change the focus point. i. In the Display Options section, you can change the Display Type to line, stick, space, primary structure, secondary structure, etc. We are going to select Space. ii. iii. iv. You can overlay the structure with a sequence conservation heat map by selecting Sequence conservation computed using all sequences in the Highlight Sequence Conservation section. This structure is obtained from A/turkey/Italy/214845/02 and the position numbering in our meta-cats analysis is the same as this strain. Now select Chain A and type in 195, 235 in the Highlight by SWISS-PROT Position section to highlight these binding determinants on the structure. G195V and Q235L could increase the binding of avian H5 and H7 viruses to human-type receptors (Yamada, 2006; Srinivasan, 2013). Alternatively, you can highlight the Sequence Feature of receptor binding determinants by checking Influenza A_H7_determinants-of-receptor-binding_194(5) from the functional dropdown list in the Highlight Sequence Features section. Variant position 235 is also located within an experimentally determined epitope. Highlight this epitope on the structure by checking Influenza A_H7_experimentallydetermined-epitope_226(14) from the epitope dropdown list in the Highlight Sequence Features section. v. Click Spin to view the structure spinning. Then click Rock to rock the structure back and forth. vi. The custom highlighted protein structure can be downloaded as an image by clicking Save View As Image beneath the image, or a 3D movie of either a spinning structure or a rocking structure by clicking Generate Video. 17

15 ome 3D Protein Structure Search Results Protein Structure Viewer (1TI8) DISPLAY OPTIONS: These options control the general appearance of the protein structure in the viewer. Display Type: Space Zoom: 100% Spin Details Rock Angle: 12 Speed: 10 Pause: 0.0 HIGHLIGHT SEQUENCE CONSERVATION Overlay the structure with a sequence conservation "heat map" (see SOP for details). Blue represents conserved regions and red represents non-conserved regions Show: Turn off sequence conservation display HIGHLIGHT LIGANDS Highlight Ligands in Reset View Save View As Image HIGHLIGHT EPITOPES Highlight epitopes on the structure in. First, select an epitope type from the list. Then check epitopes to highlight. Flash epitope during rocking Clear Description: H7 HAEMAGGLUTININ Strain Name: A/turkey/Italy/214845/02 PDB Link: 1TI8 JMOL COMMAND LINE: Advanced users can enter a JMol command directly using JMol Interactive Script. Type JMol command here. Run SAVE/RESTORE STATE You can save the current state of the image (Display Options, Zoom, Highlight Epitopes, etc.) for later use. Click the Show JMol Console button below, click State, and copy/paste/save the contents of the upper panel into any text document. When you reopen this PDB file in a later session, you can restore the image to the saved state. Click the Show JMol Console button, paste the text into the lower panel, and click Execute. Show JMol Console GENERATE 3D MOVIE Generate a mp4/avi movie, starting with the rendering in above window and spinning around Y axis by 360 deg. It takes about 30 seconds to generate the movie. Spin Rock Generate mp4 movie (default=avi) Generate Video HIGHLIGHT BY SWISS-PROT POSITION Highlight in in this structure corresponding to defined SwissProt positions. With the dropdown, select a chain and view the positions present in this file. Enter one or more comma-delimited positions (15,30) or a range (15-30) then click Highlight. If no chain is selected, these positions will be highlighted on all chains which include them. Select Chain: A - Q6GYW3 ( ,235 Highlight Clear PDB Sequence/Structure Details HIGHLIGHT SEQUENCE FEATURES Highlight sequence features on the structures in epitope Clear Sequence Feature Feature Name Sequence Feature Positions Influenza A_H7_experimentally- Influenza A_H7_SF determined- epitope_340(14) Influenza A_H7_experimentally- Influenza A_H7_SF determined- References Noronha JM, et al. Influenza virus sequence feature variant type analysis: evidence of a role for NS1 in influenza virus host range restriction. J Virol May;86(10): doi: /JVI PMID: Picket BE, et al. Metadata-driven Comparative Analysis Tool for Sequences (meta-cats): an Automated Process for Identifying Significant Sequence Variations Dependent on Differences in Viral Metadata. Virology, 2013, 447(1-2): doi: /j.virol PMID: Sorrel EM, et al. Minimal molecular constraints for respiratory droplet transmission of an avianhuman H9N2 influenza A virus. Proc Natl Acad Sci U S A May 5;106(18): doi: /pnas PMID: Srinivasan K, et al. Quantitative description of glycan-receptor binding of influenza a virus h7 hemagglutinin. PLoS One. 2013;8(2):e doi: /journal.pone PMID: Yamada S, et al. Haemagglutinin mutations responsible for the binding of H5N1 influenza A viruses to human-type receptors. Nature Nov 16;444(7117): PMID: