Discussion
The field of phylogenomics relies on a broad range of efficient software
for many operations in a typical workflow (e.g., raw read assembly, tree
inference), but other aspects of standard analysis pipelines are
neglected. We developed SEGUL to fill a need for software that can
manipulate alignment files and calculate summary statistics in a manner
suitable for computational experts and novices alike. SEGUL offers
consistently fast operations, low memory demands, and efficient CPU use
across all supported features. The convenient features of SEGUL (e.g.,
GUI support, safe command line application) offer an easy introduction
for beginners with no cost to performance for advanced users.
There is an increasing trend of relaying to containerized applications
(i.e., using Docker and Singularity) to solve dependency issues, but
this fix makes it harder for computing novices to study phylogenomics.
We take a different approach by minimizing dependencies to simplify
installation. We add GUI support to further aid evolutionary biologists
who prefer the convenience of a GUI, while retaining efficiency and
reproducibility. Using GUI version may also be more convenient to
perform the tasks on small datasets (e.g., Sanger sequencing datasets),
and for introductory teaching situations.
SEGUL support for mobile devices enables a new way to perform
phylogenomic data manipulation and summary statistics. Our application
facilitates teaching phylogenomics to students who use mobile devices
(e.g., tablets) as their primary computers. Many authors publish their
alignments (e.g., Jarvis et al., 2014; Oliveros et al., 2019,
https://github.com/roblanf/BenchmarkAlignments) and phylogenetic tree
estimation can be conducted on the web (DeSalle et al., 2020). Our
application allows the entire process from alignment inspection and
concatenation to phylogenetic tree estimation to be conducted from a
mobile device. Mobile devices are becoming more powerful, with some
tablets, such as Appleās iPad Pro and iPad Air, using the same
processors as their laptop and desktop counterparts. Future
implementations to exploit mobile device strengths, such as touch input,
would enhance the benefits of mobile applications for phylogenomic data
manipulations.
We achieve a cross-platform, high-performance application by writing our
GUI code in Flutter and handling expensive computation in Rust. Our
implementation of this approach in the neglected aspects of typical
phylogenomic workflows provides proof of concept in the development and
scaling of high-performance applications with GUI interaction on desktop
and mobile operating systems. Extension of our approach to other aspects
of phylogenomic workflows would pave the way for more user-friendly
software to study phylogenomics and enable teaching students with
limited access to computational power.