The Scottish Bioinformatics Research Network was established in 2005 to promote the development of Bioinformatics research in Scotland. It is funded by a Scottish Funding Council Strategic Research Development Grant (SRDG) for four years from Summer 2005.

SBRN funds a total of eight posts at four Scottish bioinformatics research groups, with two posts each at the Scottish Crop Research Institute, the Barton Group at the University of Dundee, the Edinburgh Centre for Bioinformatics, and the Bioinformatics Research Centre at the University of Glasgow.

The posts at SCRI support not only the bioinformatics needs of SCRI, but also those of the other Major Research Providers (MRPs) such as the Rowett Research Institute in Aberdeen, and the Moredun Research Institute in Edinburgh. The SCRI posts are held by Iain Milne and Micha Bayer. 

SBRN is involved in the development of several applications and tools also listed within the Bioinformatics subarea of the SCRI website, however additional work is also carried out with less of a public face.

Workflow analyses

We have been evaluating workflow engines that could be used to execute a variety of different analytical workflows at SCRI, e.g. DNA sequence or microarray data analysis pipelines. As part of this we have compared the Taverna workflow engine and the GenePattern platform, with the latter being chosen as our production system. We have since deployed GenePattern on the head node of our compute cluster where GenePattern jobs can be executed on the cluster through its interface with Sun Grid Engine, the cluster's job scheduling system.

BASE Microarray LIMS system

SCRI uses microarray technology for numerous purposes and there has been a need for some time to store and manage the resulting data in a Laboratory Information Management System. Due to cost considerations the SBRN team at SCRI have been exploring several free options, including maxD, local MIAMExpress and BASE, finally settling on BASE.

In using BASE we identified several possible extensions that would allow easier data entry for large batches of data, and decided to make use of the BASE extension mechanism which enables developers to build their own components and install them into BASE as "plugins", so far producing code for a batch data importer and an interface to importing data from Tab2MAGE files - a popular MIAME compliant data format.

HPC Cluster rebuild

The SCRI compute cluster provides high-throughput computing facilities for users from both SCRI and BioSS. As part of a recent rebuild of the cluster, SBRN staff worked in collaboration with IT to redesign, rebuild, and extend the cluster into a homogeneous system, harmonising the operating systems across all nodes. The cluster in its current form provides end-users with access to 70 Intel Xeon CPUs, almost 100GB of accessible memory, and well over 3TB of shared storage.

Server virtualization

As part of a drive to reorganise the bioinformatics group’s Linux servers we have converted most of our physical servers into virtual machines running on a high spec server with 2 quad core processors, 16GB of RAM and 1.5TB of storage. As of Summer 2008, this single machine is providing access to over 10 virtual machines, running a combination of web servers, database servers and development/testing environments. In addition to electricity and space savings we now also have a smaller administrational overhead and better failover protection.

SQL Data Services Generator

A growing number of applications in biological data analysis now require access to live data sources at runtime. Traditionally this has been achieved by means of direct connections into a live database but this introduces a number of complications such as the breakability of client applications as a result of changes in the underlying database schema, credential administration and security risks due to the requirement of open ports on the database server. One solution to these problems is the introduction of an additional component layer in the form of a SOAP-based (Simple Object Access Protocol) web service. A service complying with interoperability standards can be invoked from clients written in any programming language and executing on any operating system, thus providing genuine language and platform independence. 

As part of SBRN we have developed the Data Service Generator (DSG), a simple, lightweight, system that fully automates the generation of interoperable data web services from a set of SQL queries, resulting in function calls that map to predefined (“canned”) queries to the database. DSG is built entirely on mature, widely-used open source Java packages from the Apache Software Foundation (Tomcat, Axis, Ant, Log4j) and can be easily deployed and configured.

Support for MRPs

As part of our commitment to the other MRPs of the Scottish Government we currently also support bioinformatics work at the Moredun Research Institute, Edinburgh and the Rowett Research Institute, Aberdeen. Our support has been mostly focused on BASE setup, administration and plugin development, microarray workflows using GenePattern and cluster administration with Sun Grid Engine. We have made a conscious effort to maximise the overlap in the technologies used at SCRI and the other institutes, to minimise wastage of resources.

Training courses

We have provided a training course at SCRI to teach users version control with Subversion, a tool we use daily and which has been invaluable to us and attendees of our course since. We also hosted the first SBRN developers’ workshop in August 2006 which was focused on web service technologies and turned out to be a successful and well received event.