What's new?

Latest release note

2.0.21.3

• Data (Current Expression Atlas):
  • Ensembl 72 / Ensembl genomes 19 gene annotations and array design probe set mappings
  • EFO release 2.36
  • 2841 experiments (including 146 new or re-curated and re-loaded experiments)

• Features (Current Expression Atlas):
  • On each experiment page, button reveals a link to Expression Atlas Linked Data. For an example E-GEOD-5258 contains a link to http://rdf.ebi.ac.uk/resource/atlas/E-GEOD-5258.
  • As of this release, the REST API interface is deprecated. It will be replaced by a new interface when the new Expression Atlas goes live in Q4 2013. We urge all users of our current API to contact us via mailto:atlas-feedback@ebi.ac.uk so that we are able to consider their requirements when designing the new REST API. We would like to apologise for any inconvenience the transition to the new REST API causes.

• Data (New Expression Atlas):
  • Ensembl 72 / Ensembl genomes 19 gene annotations and array design probe set mappings
  • Reactome 44 gene annotations
  • miRBase 20 release microRNA annotations and design element mappings
  • Over 100 experiments including 10 RNA-seq experiments and the first microRNA microarray experiment, http://www-test.ebi.ac.uk/gxa/experiments/E-TABM-713.

• Features (New Expression Atlas):
  • http://www-test.ebi.ac.uk/gxa/experiments now shows all experiments loaded into the new Expression Atlas.
  • The help page now contains help for both baseline and differential expression experiments.
  • The gene page, e.g. DIN10, now provides referential information about a gene, as well as baseline and differential expression sections.
  • Similar pages are available for gene sets, e.g. REACTOME pathway Activation of the mRNA upon binding of the cap-binding complex and eIFs, and subsequent binding to 43S and proteins, e.g. olfactory receptor

Downloadable data and software information is available from github.

2.0.20

• Data:
  • Ensembl 71 gene annotations and array design probe set mappings
  • Reactome 44 gene annotations
  • EFO release 2.35
  • 2828 experiments (including 39 new or re-curated and re-loaded experiments)
  • As part of our ongoing effort to improve the accuracy of the differential expression calls made in Atlas, we have identified a further circa 400 experiments that cannot provide contrasts with sufficient amount of replicates per 'test' and 'reference' sample groups (where a contrast tests which genes are differentially expressed between a 'test' and a 'reference' condition). All experiments that could not be re-curated to satisfy the minimum replicates criterion were subsequently removed.
  • Expression Atlas no longer contains any RNA-seq experiments analyzed by ArrayExpressHTS pipeline. All analyzed RNA-seq experiments that satisfy our quality control criteria can be viewed within the new Expression Atlas interface instead (please see for more details below).

• Features (New Expression Atlas Interface):
  • Expression Atlas (baseline expression) now provides an breakdown per transcript for each gene-factor value cell in the results heatmap. For example, click on any non-empty cell in Illumina Body Map experiment to view a pie chart showing an expression level breakdown for up to three top expressed transcripts (plus a slice for the remaining transcripts, if any), e.g.
    
  • All Expression Atlas experiments are now searchable by REACTOME pathway accessions and descriptions. For example, view query for REACT_1619 (Death Receptor Signalling, Homo sapiens) in Illumina Body Map. Expression Atlas covers almost all known genes for this pathway (14 genes matched in Ensembl). The above example highlights leukocyte-specific expression of most genes in REACT_1619, its "tissue fingerprint". This is corroborated in literature, e.g. Death receptor signaling and autoimmunity.
  • Various types of differential expression experiments are now available in the new Expression Atlas interface. We don't yet have a unified search interface, but we would be very grateful for any feedback on the user interface and data provided for each of the experiments below:
    • Transcription profiling by array of Drosophila melanogaster CDK8 and Cyclin C homozygous mutants, determined using Affymetrix GeneChip Drosophila Genome 2.0 Array (1-colour microarray)
    • Caco-2 cells: cultured in conventional vs apical anaerobic conditions (2-colour microarray, Homo Sapiens)
    • RNA-seq of Arabidopsis mutants with defects in long-non-coding-RNA-mediated transcriptional silencing
    • RNA-seq of mouse spinal cord expressing wild type human TDP-43
    • RNA-seq of pancreatic islets from pregnant mice
    • RNA-seq of vomeronasal tissue from adult male and female mice
  • Note that for the above experiments when you click on:
    • a gene-contrast cell in the results heatmap you can see the full data behind this differential expression call, e.g.
      
    • a button, shown next to the contrast label in the header of the results heatmap for default False Discovery cutoff only, an MA plot appears for that contrast, relating average expression intensity, fold change between the 'test' and the 'reference' part of that contrast, and the differentially expressed genes (shown as red dots in the plot), e.g.
      
  • Full experimental data is now available for downloading for all differential experiments (see download buttons in the right-top corner of each experiment page)
    • Normalized data for 1-colour microarray experiments
    • Fold change data for 2-colour microarray experiments
    • Raw expression counts for RNA-seq experiments
    • Full analytics data for all differential experiments

Downloadable data and software information is available from github.

2.0.19.1

• Data:
  • Ensembl 70 annotations and array design probe set mappings
  • EFO release 2.33
  • 3168 experiments (including 68 new or re-curated and re-loaded experiments)
  • Please note that as part of our ongoing effort to improve the accuracy of the differential expression calls made in Atlas, we have also unloaded over the remaining 400 experiments that did not satisfy the 'at least three replicates per each factor value' criterion. This criterion is now also used for all new experiments being loaded into Atlas.

• Features:
  • Bug fix for pagination on heatmap results page - for queries containing apostrophy
  • Bug fix to experiment page results - affecting both best design elements table in the default experiment view and analytics downloads. Previously, for each design element only one best analytic was returned - across all factors and their values. Now, for each design element the best analytic is returned for each factor/value. This fix increases the amount of analytics data returned dramatically.
  • In view of the above bug fix, the analytics download functionality on the experiment page had to be memory-optimized. Namely, for very large experiments the data in the downloaded analytics zip file is split into chunks.

• News:
  • We have released our new Baseline Expression Atlas prototype, that aims primarily at representing information about which gene product is present (and at what abundance) in which normal condition (e.g., tissue, cell type). Please check it out - we would love to get your feedback!
  • We have also started the process of careful assessment of experimental intent and identification of contrasts for all existing experiments in Gene Expression Atlas, in preparation for their inclusion in the new contrast-based Differential Expression Atlas prototype (to be released later in 2013 - c.f. 'Future Plans' section below.)

Downloadable data and software information is available from https://github.com/gxa/gxa/wiki/Release-Notes.

2.0.18

• Data:
  • Ensembl 69 annotations and array design probe set mappings
  • EFO release 2.31
  • 3596 experiments (including 80 new or re-curated and re-loaded experiments)

• Features:
  • The links accessible on the right-hand side of the experiment page are now also accessible programmatically, e.g.
    • http://www.ebi.ac.uk/gxa/directDownload/experimentAnalytics?eacc=E-MEXP-3566
    • http://www.ebi.ac.uk/gxa/directDownload/experimentExpressions?eacc=E-MEXP-3566
    • http://www.ebi.ac.uk/gxa/directDownload/experimentDesign?eacc=E-MEXP-3566

Downloadable data and software information is available from https://github.com/gxa/gxa/wiki/Release-Notes.

Previous releases

2.0.17

• Data:
  • Ensembl 69 annotations and array design probe set mappings
  • EFO release 2.29
  • 3558 experiments (including 261 new or re-curated and re-loaded experiments)

• Features:
  • This mainly data release contains some robustness fixes for the Atlas Distributed Annotation Source (DAS) service.

2.0.16

• Data:
  • Ensembl 68 annotations and array design probe set mappings
  • EFO release 2.27
  • 3384 experiments (including 98 new or re-curated and re-loaded experiments)

• Features:
  • We restricted our statistics calculation to factors with at least two replicates per each value (as reported in http://nar.oxfordjournals.org/content/38/suppl_1/D690.full). In previous releases, incorrectly, one replicate in at least one value (rather than all values) per factor was stipulated. This fix aims to improve the accuracy of the differential expression calls made in Atlas.
  • Minor bug fixes to autocomplete, search and feedback mechanisms

2.0.15

• Data:
  • Ensembl 67 annotations and array design probe set mappings
  • EFO release 2.25
  • 3476 experiments

• Features:
  • New search for experiments functionality e.g. http://www.ebi.ac.uk/gxa/experiment/index.html?q=homo+sapiens

2.0.14

• Data:
  • Ensembl 67 annotations and array design probe set mappings
  • EFO release 2.24
  • 3364 experiments (including 262 new or re-curated and re-loaded experiments)
  • Affymetrix microarray data is now pre-processed with the oligo package from BioConductor (Carvalho and Irizarry, 2010; http://bioinformatics.oxfordjournals.org/content/26/19/2363.full). This package is better able to interpret and process data from arrays such as the Gene and Exon ST arrays from Affymetrix. 48 experiments using these arrays have been reloaded:
    • 2 Human Gene 1.0 ST array experiments
    • 43 Mouse Gene 1.0 ST array experiments
    • 3 Rat Exon 1.0 ST array experiments

• Features:
  • New links for downloading experiment design, raw expression values and differential expression analytics on the experiment page, e.g. http://www.ebi.ac.uk/gxa/experiment/E-AFMX-5 (see right-hand side)
  • Re-designed interface, including gene conditions autocomplete, for searching the experiment for best differential expression analytics, e.g. http://www.ebi.ac.uk/gxa/experiment/E-AFMX-5 (see below the plot)
  • Bug fixes to gene condition autocomplete on the man Atlas search page, to the heatmap search result display and experiment page plots.

2.0.13

• Data:
  • Ensembl 66 annotations and array design probe set mappings
  • EFO release 2.23
  • 371 new or re-curated and re-loaded experiments. This is due to an on-going major effort in Atlas data clean-up, including:
    • merging similar factors, sample characteristics and their values
    • merging compound treatments, doses and units into a single value for factor ‘compound treatment’ – to make intra-experiment expression comparisons easier (e.g. http://www.ebi.ac.uk/gxa/experiment/E-TABM-12). All newly loaded units are now part of EFO ontology – in order to remove ambiguity in data interpretation.

• Features:
  • New links for tab-delimited experiment analytics and experiment design data download. see 'Export to TSV' links on e.g.
    • http://www.ebi.ac.uk/gxa/experiment/E-MEXP-3493 and
    • http://www.ebi.ac.uk/gxa/experimentDesign/E-MEXP-3493
  • Simplifications to the Condition autocomplete functionality in both the simple and the advanced Atlas search interfaces
  • Links to EMAGE database (http://www.emouseatlas.org/emage) for Mouse genes
  • Bug fixes to the Atlas API output

2.0.12

• Data:
  • Contains 443 new or re-curated and re-loaded experiments. This is due to an on-going major effort in Atlas data clean-up, including:
    • merging similar factors, sample characteristics and their values
    • merging compound treatments, doses and units into a single value for factor ‘compound treatment’ – to make intra-experiment expression comparisons easier (e.g. http://www.ebi.ac.uk/gxa/experiment/E-TABM-12). All newly loaded units are now part of EFOontology – to remove ambiguity in data interpretation.
  • Contains 53 RNA-seq experiments
  • Updated with Ensembl 66 annotations and design element mappings

• Features:
  • Allows searching using gene signatures from GeneSigDB Data Release 4 (http://compbio.dfci.harvard.edu/genesigdb/)
  • Includes bug fixes to experiment page, main search page and minor improvements to the API functionality.

2.0.11 and before

• Please see: https://github.com/gxa/gxa/wiki/Release-Notes

Future Plans

Baseline Expression Atlas Prototype

Majority of our effort from Autumn 2012 onwards has been spent on putting together the Baseline Atlas prototype (still under heavy development), due to be released in the first quarter of 2013. Baseline Atlas aims primarily at representing information about which gene product is present (and at what abundance) in which normal condition (e.g., tissue, cell type). In the first instance Baseline Atlas will consist of individual well understood and curated RNA-Seq experiments, such as: Illumina Body Map Transcriptome of DBAxC57BL/6J mice RNA-seq of long coding and long non coding RNAs from ENCODE cell lines RNA-seq of 6 tissues from 10 species to investigate the evolution of gene expression levels in mammalian organs Baseline Atlas will allow the user to explore gene signatures for normal conditions and combinations of conditions, at different levels abundance cutoffs (where all expression on or below a chosen cutoff will be considered background).

Differential Expression Atlas for Individual Contrasts

Multi-omics Atlas

We are working closely with the EBI's PRIDE and Metabolights teams on idientifying data sets, statistical analysis methods and visual integration methods aiming to bring transcriptomics, proteomics and metabolomics data into a single searchable user interface. We will be providing more detail as our work progresses.