Biomedical Engineering Reference
In-Depth Information
procedure, together with detailed documentation. To install the standalone
Atlas, two fi les are downloaded - a web application and a data archive.
The data archive contains an installer, which unpacks the fi les and loads
them automatically into the specifi ed database. The web application
contains an administration interface with confi guration options and data
loading interface. We can release incremental updates to the data archive,
and update an Atlas installation without re-installing the database.
9.3.4 Creating a new administration interface
for loading data
A completely new administration interface was developed, as shown in
the fi gure below. This interface allows the Atlas administrator to load
and unload experiments, update data and analytics, and manage various
confi gurable options of the software, for example custom branding
templates and external database/cross-reference links.
9.3.5 Future enhancements
As Atlas matures into a larger platform for transcriptomics data, we have
started to expand its scope from microarrays towards other 'omics data
types - fi rst of all, high-throughput sequencing-based, or next-generation
sequencing (NGS), transcriptomics. To this end, we developed a pipeline,
ArrayExpressHTS [14], for processing NGS data sets from raw reads to
quantifi ed transcript isoform expressions, and have been working on
integrating it into the Atlas. NGS data, RNA-Seq, and ChIP-Seq will be a
key focus for future Atlas developments.
The amount of data in the public Atlas has grown more than sixfold
since its start in 2009 and we are seeing that for a growing number of
experiments our statistical approach may not be best suited. We must
turn our attention now to method development: in particular, we are
gradually switching to using raw data wherever possible and reprocessing
these data for best standardization and comparability, and we are
working with other groups to incorporate advanced gene set-based
statistics into our pipeline.
Lower costs and technological barriers have made possible two
comparatively new kinds of experiments: large meta-analysis experiments,
such as the global map of human gene expression (Atlas experiment ID:
E-MTAB-62) and large reference experiments such as the Illumina Body
Search WWH ::
Custom Search