000K utf8 0100 1028881673 1100 $c2018 1500 eng 2050 urn:nbn:de:gbv:27-dbt-20180814-1012346 2051 10.22032/dbt.35145 3000 Saraiva, Joao 4000 Distinguishing fungal from bacterial infection$da mixed integer linear programming approach [Saraiva, Joao] 4060 129 Seiten 4209 The immune system is responsible for protecting the host from infections. In healthy individuals, this system is generally able to fight and clear any pathogen it encounters. Blood stream infections can be caused by several pathogens such as viruses, fungi and bacteria. Delivery of appropriate treatment requires rapid identification of the invading pathogen. The use of in situ experiments attempts to identify pathogen specific immune responses but these often lead to heterogeneous biomarkers due to the high variability in methods and materials used. Support Vector Machines (SVMs) allow using gene expression patterns to discriminate between two types of infection. Comparing gene lists from independent studies shows a high degree of inconsistency. To produce consistent gene signatures, capable of discriminating fungal from bacterial infection, SVMs using Mixed Integer Linear Programming (MILP) were employed allowing the combination of classifiers using different datasets. Employing this method demonstrated the improvement in consistency of the produced gene signatures that distinguished fungal from bacterial infections irrespective of the type of the leukocyte or the experimental setup. The produced biomarker list showed an increase in consistency of 42% when compared to single classifiers and predicted the infecting pathogen on an unseen dataset with an average accuracy of 87%. Restricting the analysis to datasets comprised of peripheral blood mononuclear cells and monocytes, showed an enrichment of genes from the lysosome pathway that was not shown when using independent classifiers. Moreover, the results suggested that the lysosome pathway is specifically induced in monocytes. In conclusion, the combined classifier approach increased the consistency of the gene signatures, compared to single classifiers and "unmasked" the monocyte-specific expression profile for fungal infections. 4950 https://doi.org/10.22032/dbt.35145$xR$3Volltext$534 4950 https://nbn-resolving.org/urn:nbn:de:gbv:27-dbt-20180814-1012346$xR$3Volltext$534 4961 http://uri.gbv.de/document/gvk:ppn:1028881673 5051 610 5550 Detektion 5550 Genexpression 5550 Immunologie 5550 Immunozyt 5550 Infektion 5550 Infektionskrankheit 5550 Medizin 5550 parasitäre Krankheiten