In order to support DArT technology several software tools and systems have been developed and implemented. Because of the very specific needs and novelty of DArT, most of the system components have been developed "in-house". The system has four main components:

• Laboratory management (DArTdb);
• Genotypic data analysis (DArTsoft);
• Breeding software;
• Additional tools for distance analysis, score merging, creating genetic maps and association analysis.

DArTsoft

DArTsoft is a purpose build standalone application. It has been developed with high volume data analysis in mind, so the whole data extraction and analytical process is higly automated. There are two major components in the package:

• microarray image analysis
• polymorphism and scoring analysis

Microarray image analysis can be performed on local TIF files using various formats of printmaps. Spots on images are localized in automatically (see example control image [295 kb]). The results are local data files containing relative hybridization intensities, raw intensities, background values and over twenty spot quality parameters. It also produces several control images, which can be used to visualize analysis results.

The algorithm for image segmentation uses a very efficient 'seeded region growing' approach to locate spots. It allows to find spots accurately even for irregular spots of various sizes. The algorithm also corrects rotated images when required.

The same algorithm has been integrated into DArTdb and interfaced there, so that images stored in the database can be analyzed and the data stored, directly in the database.

Polymorphism and scoring analysis can be performed in both local and database modes. Local mode uses the results of local image analysis. Database mode requires connection to DArTdb and uses the database tables where raw image results are stored. This ensures fully automatic and correct clone and sample tracking. The database mode is absolutely critical for handling big volumes of data.

Polymorphism analysis identifies polymorphic markers and scores them. For each marker, the software classifies relative hybridisation intensities on all slides of the population (or experiment) analysed. Scoring of the markers is done in a dominant (bimodal distribution of intensities), or hemi-dominant (trimodal distribution) manner. The program reports scores, confidence estimates, call rate and PIC values. Our technology is quite unique in its quantitative approach to evaluating marker quality and the confidence associated with each individual genotype call (score). DArTsoft computes quality and confidence estimates, and these features will be critical to improve the performance of downstream data-analysis software packages.

DArTsoft also contains a function for importing results into Excel for easy data viewing and filtering.

DArTsoft will be available to members of the DArT network, and possibly outside of this framework.

DArTdb

DArTdb is our Laboratory Information Management System (LIMS). It is a purpose build storage, data management and analysis system, which has been designed to follow exactly the technology workflow. Thus, DArTdb records all the data the technology relies on and provides a set of tools and utilities to manage them. The processes covered by DArTdb include technology development as well as the provision of the genotyping service.

At the design stage of the system the following constrains had to be taken into accout:

• High-throughput of the technology produces big quantities of data;
• Risks of sample and clone tracking errors has to be eliminated through integration of the laboratory hardware into the system;
• Identification of the sample and clone plates as well as clone assays has to be seemless;
• Access to the data has to be easy, quick and secure;

Following the above guidelines DArTdb is an access-controlled application with three different levels of user's permission to access database functions and data - administrators, power users and users. This access control along with a multilayer backup/archive system assures high standard of data security.

DArTdb is closely integrated with our on-line ordering system as well as with DArTsoft - our analytical software. This assures direct access to any data in a secure and integrated way eliminating potential problems during data transfers and conversions.

Barcoding system of all plates and clone assays assures precise identification of all clones, samples and hybridizations.

In general DArTdb has been designed to perform the following functions:

• System administration - which includes administration of the system users, management of laboratory hardware, management of controlled vocabularies (e.g. types of plates) and so on;
• Service provision - management of clients, products and orders - which includes workflow tracking of each order;
• Annotation of DNA source names - organisms, species, genotypes and tissues;
• Management of DNA samples - which includes tracking plates using barcodes, DNA quality information and annotating samples using source names;
• Management of clone libraries - which includes clone sources and naming, tracking library plates using barcodes, clone tracking and identification;
• Definition of microarray designs and printing clone assays - which includes tracking assays using barcodes and mapping of each assay to the clone;
• Defninition of hybridisation experiments - with tracking of the experiment process;
• Raw data automatic import - as a result of the experiment TIF images are produced and they consist of raw results;
• Image data extraction - using DArTsoft image module integrated directly into DArTdb;
• Additional utility functions - e.g. barcode search, hybridization assistant and others, producing reports helpful in everyday work;

From the technical perspective DArTdb has been designed in a client-server architecture. In order to maintain high security and data integrity, backend of the system has been based on a relational database. Access to DArTdb application has been implemented thtough the web browser environment, which assures easy access to the data, without a need to preinstall or configure any additional components. To maximise stability and minimise cost, DArTdb has been entirely built using Open Source components (using the so-called LAMP architecture: Linux, Apache, MySQL and Perl).

DArTdb will be available to members of the DArT network.

Breeding software

We fully recognise the importance of information technologies, and breeding software in particular, for the successful application of whole-genome profiling. We also recognise that some of our clients may not have sufficient access to software in this critical area. We have therefore established a partnership with the leading breeding software company, GeneFlow Inc., to facilitate access to their software for breeders in Australia. This software can efficiently store, process, filter and visualize both phenotypic variation and marker information, making breeding with markers more efficient.