Opal 2.x

OBiBa’s core database application for biobanks

About this presentation

  • Hit the left/right arrow to browse to the main sections
  • Hit the up/down arrow to see the slides in each section
  • Hit the "escape" key to see all the slides


You can Try it !

  • username: administrator
  • password: password

What's new

  • Projects
  • Administration
  • R
  • DataSHIELD

Projects

A project is a repository of data and dictionaries.
In a project, data are imported, exported, analysed, transformed and shared. 

Try it

Project Tables

Formerly known as a datasource.

Project Tables

A table has a dictionary (variables) and values.
  • Each variable can be edited individually,
  • There is a summary for each value type.


Project Tables

Derived variable scripts versioning.

Project Files

A project has its own folder to allow collaborative work.

Project Reports

Use R and Markdown for reporting.

Project Tasks

Follow import, export, copy, report tasks.

Project Permissions

Manage project permissions in one page.

Project Administration

Description, data storage, key store...

Administration

Administration was improved to make Opal more dynamic.
  • Data accessusers, ...
  • Data managementidentifiers, ...
  • Data analysis: R, ...
  • System: databases, JVM...


Users

  • Users and groups can be added from the web interface,
  • Users can be authenticated by password or certificate.

Identifiers Mappings

  • Identifiers of any entity type can be mapped,
  • Easier identifiers management.

R Service

R server can be controlled, even if it runs on a different machine.


Databases

  • Easier database management,
  • MongoDB for large datasets.


Java Virtual Machine

  • Memory usage
  • Threads
  • Garbage collector activity


R

Opal and R in 5 lines.
# Login in opal
library(opal)o <- opal.login('administrator', 'password', 'http://localhost:8080')
# Assign a table to a data.frameopal.assign(o,'D','mica_demo.FNAC')
# Analyse data in the remote R sessionopal.execute(o,"summary(D)")

# Get data into the local R sessionD <- opal.execute(o,"D")
# Terminate remote R sessionopal.logout(o)

DataSHIELD

DataSHIELD process is parallelized.

Was sequential

  1. R command is executed on Opal 1
  2. ...
  3. R command is executed on Opal n
  4. Results are combined
Now is parallel
  1. R command is submitted to Opal 1, ..., Opal n
  2. R command result is retrieved from Opal 1, ..., Opal n
  3. Results are combined

Resources

opal2

By Yannick Marcon