Identifying gene subsets through Gene Ontology based merging algorithms AND Assessing Handheld Computer Use by Physicians in Preventing Medication Errors in Nursing Homes
Identifying gene subsets through Gene Ontology based merging algorithms
Vicky Chen, BS, Doctoral Fellow
Abstract: Motivation: The Gene Ontology (GO) is a controlled vocabulary of biological concepts used to annotate gene products and PubMed articles. Most of the available gene summarization tools do not use semantic information in their calculation. This study develops and evaluates automatic merging algorithms that use semantic similarity and gene counts when identifying coherent gene subsets.
Results: We developed a software package that can represent the GO. Using this representation, we extended previous research that utilized semantic similarity and the information bottleneck method by incorporating a functional coherence measure to determine when to stop the merging process. Graphs produced by three variations of the merging algorithm were compared with a randomly merged graph. We found that there is not a significant difference between the three non-random merging algorithms, even though they all returned results with greater significance than the random merges.
Assessing Handheld Computer Use by Physicians in Preventing Medication Errors in Nursing Homes
Frank Ligons, BS
Abstract: This manuscript describes a survey of physicians involved in geriatric care, inquiring about their use and perceptions of mobile drug reference technology in preventing adverse drug events (ADEs). Our team is interested in how this technology may improve patient care by aiding in the prevention of ADEs in the nursing home (NH) environment. Our findings indicate that the use of mobile drug reference technology in NHs is moderately widespread. We also found that a meaningful portion of physicians believe that ADEs were prevented by this technology in the month prior to this study.
We hope our study contributes to the discussion of mobile drug reference technology as a potential intervention in reducing medication errors. We believe this work is an important prelude to prospective trials comparing medication error rates between groups that do and do not employ this technology.