Понравилась презентация – покажи это...
Patient profiling: Big Data’s impact on Healthcare Rene Kuipers Principal Consultant Big Data & Analytics @rjlkuipers
Master’s degree (1998) Biochemist by training Molecular Biology Tumorgenetics IT since 1999 Oracle specialist Database Analytics Big Data Founded Healthcare Analytics in 2014 @rjlkuipers slideshare.net/rjlkuipers1 renekuipers.wordpress.com Speaker Bio - Rene Kuipers
Medical examination personalized healthcare n=1 treatment. shrink cohorts by analyzing more data. patient-profiling
3.2 billion bases ACTG 2 TB raw data comparisons: enormous amounts of data
External vs Internal DATA
#headache We share more with our ‘followers’ than we do with our doctors…
Oracle’s efforts Big Data Appliance Capturing of non-structured data Webcrawling and capturing external data Big Data SQL Query non-structured data and structured data with a single (and well-known) language: SQL Big Data Discovery Early pattern detection
Oracle’s efforts Oracle Exadata Fastest database machine for Oracle databases Extreme query response times Extreme data reduction by means of compression Oracle Database Highly scalable Industry standard Support for extreme large datasets
VX Healthcare Analytics efforts Huvariome Web-environment for comparison of Whole Genome Sequences Built on Oracle Exadata / Oracle Database Server-side queries Clients use a webbrowser, no client-tools needed Scientific publication Stubbs et al. Journal of Clinical Bioinformatics 2012, 2:19 http://www.jclinbioinformatics.com/content/2/1/19 Offered as on-site implementation or Cloud-solution Allows for genomic patient profiling.
VX Healthcare Analytics efforts Cloud offering for Oracle Translational Research Center Combine clinical and genomic data As well as lab results Intervention data Encounter data BI front-end for cohort analyses Custom visualizations based on open-source standards Open platform: connect your own preferred analysis tools
Personalized Healthcare = Utopia ? The more we know, the more we know that we don’t know much. We need more (external) data to even come close to personalized diagnoses, let alone personal treatments. patient-profiling is promising from a prediction point of view.. Conclusion