Computer Science and Software Engineering Capstone Presentations
Fall Quarter
December 18, 2020
Marko Lakic Bazen Nega "Web-QuateXelero:
An Efficient Web-Based Motif Detection Tool" (Group Project - Student Defined) Faculty Advisor: Dr. Wooyoung Kim |
Abstract A network motif is a statistically overrepresented
subgraph pattern in a given network. Network motifs in bioinformatics play an
important role in discovering significant biological functions. They are
found as follows: the user specifies a subgraph size, a number of random
networks, and a statistical threshold (Z-Score or p-value). Then, all
subgraphs of that size are enumerated in the network, a collection of the
specified number of similarly structured random networks are generated and
the same process is repeated on those networks. Statistical analysis shows if
the frequency of any subgraph in the original network is above or below in
random networks, if the difference exceeds your threshold, it is a motif.
Although various algorithms and tools are available to detect network motifs,
most of them lack accessibility and usability. Some are command-line programs
with manuals needed to navigate while others are mere algorithms without a
released program implementing them. The most popular program, FANMOD,
utilizes one of the most efficient algorithms, ESU, and provides a graphical
user interface with motif image visualization in its results. Although, there
are some nuisances. Necessarily, one must download an executable, and there
may be system incompatibility issues. An online tool gets around this, and
there are few of them. Most of these tools use the aforementioned ESU
algorithm. QuateXelero is a faster algorithm that exclusively has a
command-line interface. We have decided to mix the benefits of these two
programs by presenting "Web-QuateXelero", an online graphical user
interface (GUI) program to detect network motifs efficiently. Experimental
results show that the site runs faster than other online tools and FANMOD. We
also have motif image visualization as an optional feature (which increases
time) and implement a backend queue for taking requests when server load is
high. |
|
Updated November 24, 2020