24th Annual Student Research Conference
Automated Functional Annotation of Maize Genes Using a Support Vector Machine
Kelsy M. Parker* and Mike J. Rodriguez
Dr. Jon Beck and Dr. Diane Janick-Buckner, Faculty Mentors
An ongoing series of investigations has identified a set of over 6,000 Zea mays genes of interest in shoot apical meristem (SAM) function and leaf primordial development. High-quality manual functional annotations of over 1,000 of these genes has resulted in a database of gene ontology (GO) numbers and enzyme information for gene products with known EC numbers (http://www.genome.jp/kegg/). These annotations are being used as the training set for an automated annotation system using a support vector machine (SVM) system. The system uses cDNA sequences with BLASTX to identify relevant articles in PubMed; SVM input attributes are harvested from the abstracts of these articles in order to predict GO terms primarily in the biological process sub-ontology. Current work is focused on improving the number and quality of attributes and on increasing the selectivity of abstracts associated with a given cDNA sequence
Keywords: SVM, maize, genetics, annotation, GO
Topic(s):Computer Science
Biology
Presentation Type: Poster
Session: 6-1
Location: Georgian Room - SUB
Time: 4:30