Wasabi functionality
What file formats are supported?
For importing phylogenetic tree and sequence data, Wasabi recognizes Fasta, Fastq, NEXUS, Phylip, Clustal, JSON, HSAML, PhyloXML, Newick and New Hampshire eXtended (NHX) file formats. The import system does not depend on file name extension and it uses a fallback procedure to retrieve sequence data from unrecognized file formats.
Loaded datasets can be exported from Wasabi to Fasta, NEXUS, Phylip, HSAML, Newick and NHX file formats. The Phylip output is in sequential format and compatible e.g. with RAxML and PAML software. To use the data with the original Phylip tools, one may have to specify the option that cuts the sequence names after 10 characters.
How can I import tree and sequence data together?
To open aligned sequences together with associated phylogenetic tree, provide all of the data files in the import dialog using one of these options:
- Drag and drop a selection of files
- Click the “Browse” button and select multiple files
- Add one or more text fields (click
) and fill the input with raw data or file URL
An option to add files to already loaded data is planned for the next version of Wasabi.
How do I browse analysis history?
The analysis history browser shows a list of stored analyses, which can be viewed step-by-step using the 
button. The settings button 
in window header provides switches for compact view mode and for displaying analysis history path. Click on the triangles (
or 
) to reveal more information about a specific analysis step (metadata or list of files, respectively).
Can I have multiple user accounts?
Multiple Wasabi user accounts can be created as needed, and resulting web addresses (in form: wasabiapp.org:8000/UserID) are used to switch between the accounts. To create additional Wasabi account, disable “Remember me on this computer” in Wasabi settings (or use incognito mode), reload Wasabi webpage and follow the prompt. Interface for multiple account management is planned for a future version of Wasabi.
Why are some internal tree nodes marked with a colored circle?
Some tree nodes may be visualized as a circle filled with specific color to highlight its properties. Hovering mouse cursor over a tree node reveals its description. Here is a summary of node colors:
- White – ancestral sequence is attached
- White (larger circle) – subtree is collapsed
- Red – modified/rearranged node
- Blue – duplication node
How does a local Wasabi installation differ from the public web server?
The local installation packages (avaliable here) include the same soure code files (and functionality) as the public server. However, the server module configuration and some interface elements of local Wasabi are tailored for offline use. Main differences from the web-based Wasabi include:
- One analysis library (user accounts are disabled)
- Sharing links are disabled (not usable offline)
- Interface for some server module settings
- Automatic installation system of Wasabi updates
The default setup of local Wasabi can be changed by modifying server module settings in “default_settings.cfg” text file. For example, to enable user accounts (e.g. for a shared Wasabi installation in a web server), edit “useraccounts = NO” to “useraccounts = YES”. All options are documented with comments.
PRANK and PAGAN aligners
What is the phylogenetic tree shown within Wasabi and how is it generated?
For PRANK an PAGAN alignments, the phylogenetic tree represents the alignment guidetree and the resulting alignment is consistent with that tree. For a new analysis (unaligned sequences, no guidetree), PRANK an PAGAN first use MAFFT to generate a quick alignment and then estimate a tree from that alignment either using Neighbor Joining or FastTree. They then use this tree as the guidetree for phylogeny-aware alignment and output it within Wasabi. If the tree would be estimated from the resulting alignment, it could look slightly different.
What happens when an alignment is updated?
Wasabi can show inferred sequences at the ancestral nodes of the phylogenetic tree. If the tree is modified (e.g. by moving a subtree to a new location), an ancestral node get deleted and a new ancestral node is created. Whereas all extant sequences are moved to the their correct positions, some ancestral sequences are missing and a small number of others have new descendants and their inferred states may not valid any more. Thus, even if the alignment itself is taken as it is, some ancestral sequences have to be re-estimated. Before doing that, branch lengths may also be re-estimated.
How do PRANK and PAGAN work and what other programs do they use?
PRANK and PAGAN use other tools for specific tasks. PRANK currently uses following other programs:
- MAFFT for initial alignment (optional)
- FastTree for guidetree estimation (optional)
- Exonerate for alignment anchoring (optional)
- RAxML for ancestral state reconstruction and branch length re-estimation (optional)
PAGAN may additionally use:
- BppDist and RAxML for guidetree estimation (optional)
- BppPhySamp for data pruning (optional)
- BppAncestor for ancestral state reconstruction (optional)
Both programs start with a quick alignment generated with MAFFT, estimate a guidetree from that alignment using the specified phylogenetic inference method and perform then the phylogeny-aware alignment. For the resulting alignment the both programs may estimate ancestral character states with RAxML/BppAncestor and combine these estimates with the self-estimated character presence/absence patterns. These ancestral sequences can then be shown within Wasabi.