Unbiased comparison of data sets
Severe Asthma is often difficult to manage and many patients are unresponsive to treatment. Furthermore, it is thought that there are many different phenotypes of asthma that are not properly understood. U-BIOPRED aims to create ‘handprints’ that identify sub-phenotypes of asthma. The handprints can then be used to better understand the disease and lead to better targeted treatments for the individual.
U-BIOPRED handprints include a wealth of diverse data on each patient and the researchers need to compare and combine them in an unbiased environment. This will permit identification of a wide range of causes and contributing factors that can lead to asthma and give a detailed view of the differences that characterize different sub phenotypes of asthma and particularly those which characterise severe asthma.
The eTRIKS solution
eTRIKS has been working with U-BIOPRED to create an environment in which the diverse data sets can be compared in an unbiased way – deploying cutting edge analytical techniques to help them in their search. Working jointly eTRIKS and U-BIOPRED have created a world leading Knowledge Portal in which the data from U-BIOPRED, include lifestyle, clinical, lipidomics, unbiased proteomics, metabolomics and immunonhistochemistry can be brought together, tracked and analysed. eTRIKS trains the U-BIOPRED scientists in the use of tranSMART using their own data – greatly increasing the relevance and retention of the training.
Using the Knowledge Portal the results of analyses can be shared within the project. With the ground breaking patient engagement model the analyses can be further shared with patient organisations to guide patient recruitment and dissemination of the results.
Together eTRIKS and U-BIOPRED have been awarded the prestigious BioIT World Award for 2014.
Working with eTRIKS
- tranSMART 1.1 instance installed
- User training using project’s own data – webinar and on-site
- Sample Explorer developed to review samples available for selected cohorts
- Data modeling and remodeling to accommodate new data and new analytical techniques and to provide an intuitive tree of data entities
- Data quality monitoring, curating and loading of adult and pediatric data
- Data types include clinical, lipidomics, unbiased proteomics, metabolomics and immunonhistochemistry
- Data loaded for dust mite and flu challenge animal studies
- Pathway data modeled to allow filtering on gene expressions from KEGG
- Data analysis support provided to U-BIOPRED scientists
- Developed and deployed standards to compare diverse data sets
- Building ECLIPSE data set to compare with U-BIOPRED data