|
Research Projects
Fall Term (September-December 2005)
1)Refactoring of Pathology Input Tool
Due to minor modifications that were made to the pathology database
tables, I refactored the Pathology Input Tool to reflect these
changes.
2)Adding New Functionality to TCT
For the primary purpose of being able to further visualize and
interpret coactivity, I refactored the TCT to support a new
functionality called the coactivity interpretation tool (CAIT).
CAIT is based on the concept of coactivity for an ontological
pair. The TCT now allows the coactivity, displacement, and average
activity indices to be displayed as a heat map by overlapping
with the original TCT plot.
3)Tree Map
My largest project for the term, the TreeMap, is a visualization
for hierarchical data. The Tree Map takes (pathology) data from
either dbZach or an input file and parses the data into a tree-like
structure. I have developed and implemented the recursive algorithms
for three layouts: Clustering, Slice-and-Dice, and Squarified.
4)Replication of Yan Sun's DRE Search
This ongoing project is to continue replicating an experiment
done by a previous post-doc, Yan Sun. The experiment involved
searching for dioxin response elements (DREs) in Human, Mouse,
and Rat genomic sequence. I have created a small Java application
that finds and reports DRE occurrences in 20 000 15 000 bp sequences
generated using a Mersenne Twister random number generator.
I have replicated some of the bioinformatics components of the
experiment, including statistical analyses in R, and analysing
the occurrence of DREs in orthologous genes requiring extensive
querying of data from Ensembl and HomoloGene.
5)Software Comments Interface
This project is a Java WebSerlvet to allow members of the lab
to submit comments about in-house software. The interface requires
the incorporation of Java’s Web Development kit, JDBC,
as well as SQL queries to insert software comments into dbZach.
Winter Term (January-April 2005)
1)Pathology Input Tool
This first project is a Java application that allows pathology
data produced in the lab to be inputted into dbZach. Implementation
required the leveraging of Swing, OJDBC, and JExcel libraries.
Through SQL queries, the Pathology Input Tool inserts pathology
lab and histology data in the form of a text file or Excel spreadsheet.
2)Toxicogenomics Correlation Tool (TCT)
This second project is a visualization tool to provide a means
of interpreting data from cross-comparison studies, such as
that for mouse and rat orthologues. This Java application reads
raw gene expression and activity data text files, calculates
the correlation data, and plots the correlation data. The TCT
supports correlation filters, and data output for user-selected
points on the plot.
3)K-Means and Improved K-Means Clustering
This third project involves two implementations of K-Means clustering
visualizations. The first implementation is essentially the
classical K-Means algorithm which supports the Euclidean, Manhattan,
Minkowski, and correlation distance metrics. The second implementation,
called the Improved K-Means, builds on the first by performing
a specified number of iterations of the classical K-means. Using
these iterations, the Improved K-Means determines how often
each gene occurs in the same cluster, and displays this information
as a linkage plot.
4)Electrophoresis gels
Throughout the Winter term, I performed many electrophoresis
gels for Jeremy’s Primer project. I also did RNA gels
for Cora, and verification of restriction enzymes for Darrell.
|
