articles were selected and marked by oncologists and research staff from the Institute for Personalized Cancer Therapy at the UT MD Anderson Cancer Center. For the selected genes, we obtained 93.6% precision for gene name disambiguation and 80.4% AUC for gene and article association. For additional 223 human genes relevant to cancer, by using the Ontology Fingerprints generated from the publications before December 20, 2009 for these genes to predict the association of these genes with papers published after 2009, we got a highest precision up to 92.7%.
Personalized cancer therapy relies on extensive knowledge of cancer genes, their variants and treatments that target these variants. While most of this knowledge can be extracted from the biomedical literature, identifying genes and their associated publications with high precision is still a daunting task, often challenged by ambiguous gene names in the text. One way to disambiguate gene name is through gene normalization - the task of mapping a named entity in text to an identifier in a database. However, many genes have multiple names or aliases, part of them share identical names, even though they are distinct genes with different functions. Developing new methods to distinguish these ambiguous gene names will significantly improve the accuracy of information retrieval and other research-enabling applications.
To overcome this hurdle, we generated a nonsupervised approach to create ontology profiles termed Ontology Fingerprints for selected genes that are relevant for personalized cancer therapy from the literature. The Ontology Fingerprint for a gene consists of a set of associated GO terms and their ancestors defined by biologists, with an enrichment p-value mapping to each term to reflect the significance of the term. We first used the ABGene/GNAT to identify gene names from the PubMed abstracts, and matched the names to the gene name or alias of known genes. The ambiguous names were then assessed by evaluating the degree to which the abstract matched the Ontology Fingerprints of the genes.
GPU-based MapReduce framework to handle big data and to improve performance. Comparing with running the program on Lonestar cluster, we can gain the same magnitude of speed when using the GPU MapReduce framework. Overall, the MapReduce framework makes execution of the program more convenient and affordable, especially on a workstation with an appropriate graphic card.