Our Unity interface
One of our team members exploring the phylogenetic forest!
Length Linearization of One Protein, P53
Our implementation of a binary tree in C#
Example of Phylogenetic Tree to Parse from EggNOG
What it does
PhyloForest helps researchers and educators by improving how we see phylogenetic trees. Strong, useful data visualization is key to new discoveries and patterns. Thanks to our product, users have a greater ability to perceive depth of trees by communicating widths rather than lengths. The length between proteins is based on actual lengths scaled to size.
How we built it
We used EggNOG to get phylogenetic trees in Newick format, then parsed them using a recursive algorithm to get the differences between the protein group in question. We connected names to IDs using the EBI (European Bioinformatics Institute) database, then took the lengths between the proteins and scaled them to size for our Unity environment. After we put together all this information, we went through an extensive integration process with Unity. We used EBI APIs for Taxon information, EggNOG gave us NCBI (National Center for Biotechnology Information) identities and structure. We could not use local NCBI lookup (as eggNOG does) due to the limitations of Virtual Reality headsets, so we used the EBI taxon lookup API instead to make the tree interactive and accurately reflect the taxon information of each species in question. Lastly, we added UI components to make the app easy to use for both educators and researchers.
Challenges we ran into
Parsing the EggNOG Newick tree was our first challenge because there was limited documentation and data sets were very large. Therefore, it was difficult to debug results, especially with the Unity interface. We also had difficulty finding a database that could connect NCBI IDs to taxon information with our VR headset. We also had to implement a binary tree structure from scratch in C#. Lastly, we had difficulty scaling the orthologs horizontally in VR, in a way that would preserve the true relationships between the species.
Accomplishments that we're proud of
The user experience is very clean and immersive, allowing anyone to visualize these orthologous groups. Furthermore, we think this occupies a unique space that intersects the fields of VR and genetics. Our features, such as depth and linearized length, would not be as cleanly implemented in a 2-dimensional model.
What we learned
We learned how to parse Newick trees, how to display a binary tree with branches dependent on certain lengths, and how to create a model that relates large amounts of data on base pair differences in DNA sequences to something that highlights these differences in an innovative way.
What's next for PhyloForest
Making the UI more intuitive so that anyone would feel comfortable using it. We would also like to display more information when you click on each ortholog in a group. We want to expand the amount of proteins people can select, and we would like to manipulate proteins by dragging branches to better identify patterns between orthologs.