How to search ArrayExpress

Browse experiments

This section describes how to search ArrayExpress for experiments of interest and how to interpret the output.

ArrayExpress can be searched using any keyword (for example RNAi or breast cancer). Search terms consisting of more than one word must be entered inside double quotes, for example "prostate cancer", otherwise the two words will be treated as independent and the search run for the two words separately (in this case 'prostate' AND 'cancer' instead of "prostate cancer"), possibly resulting in an exceedingly long list of search results.

Alternatively, you can start your search by selecting the 'Experiments' option. This will display all the experiments. You will be able to refine your search after this first step (Figure 3).

Figure 3 Browsing ArrayExpress




  1. Open ArrayExpress in a new window.
  2. Click on 'Experiments' [A].
HelpFor a full list of search options and search syntax please refer to the ArrayExpress documentation.

Browse experiments -The output view

The entire content of the ArrayExpress is now shown (Figure 4). A list of all the experiments is presented ordered by release date; the most recently released experiment is at the top of the list. You can sort the results by clicking on any of the column headings. For example, if you want to sort by species, just click on the 'Organism' column header.

Figure 4 Browse experiments -The output view



[A] The Accession column contains the unique experiment accession number. Experiment accession numbers are in the format of E-XXXX-n, where XXXX is a code for the type of submission utilised (for example MTAB, if submitted using MAGE-TAB or MEXP if submitted using MIAMExpress) and n is a number assigned to each experiment. Experiments imported from GEO have accession numbers in the format of E-GEOD-n. The original GEO series identifier is kept as a secondary accession number and can also be used to search ArrayExpress.

[B] The Title column contains the experiment description as supplied by the submitter.

[C] The Type column provides information on the experiment design.

[D] The Organism column provides information on the species used.

[E] The Assays column gives you information on the number of assays in the experiment.

[F] The Released column displays the experiment release date.

[G] The Raw/Processed/Atlas columns  contain information on the type of data (raw and/or processed) available for each experiment. Moreover, it will tell you if the experiment is also loaded in the Atlas. For raw HTS data a link to the European Nucleotide Archive (ENA) entry is provided.

[H] Save the output. The list retrieved can saved in Tab-delimited format or exported in XML format. An RSS feed is also available for updates on the latest datasets loaded.


If you need further infromation on the Expression Atlas, please refer to the 'Expression Atlas' course.

Refine your search

Once you have a list of all the experiments present in ArrayExpress, you can add terms to the search box to refine your search and retrieve only experiments of interest (Figure 5).

Figure 5 Refine your search





[A] You can use the experiment search box to refine your query. To search for a particular experiment of interest start typing in the search box. As you type, the auto-complete function will try to complete the term by suggesting possible matching terms. To add one of the suggested terms to the search box click on it.

[B] The experiment search box uses a controlled terminology: the Experimental Factor Ontology (EFO). This automatically expands your searches to include synonyms and to suggest related terms. If you need any further information on EFO, the course on the Expression Atlas has a page describing it in the 'What is the Atlas' section.

[C] You can use the available filters to reduce the number of experiments retrieved. You can select an organism of interest and/or the technology used to perform the experiment (for example microarray or HTS). Moreover, you can select a molecule of interest, such as DNA, RNA or protein.




3. In the experiment search box type a keyword of interest and select the matching suggestion. In our example, we were interested in 'chronic myelogenous leukemia' so we typed this in the search box and selected the matching suggestion. To complete the search click the 'Search' button.

4. Select the organism of interest from the drop down menu [C]. In our example we were interested in Homo sapiens. To add this filter click the 'Filter' button.

InformationBe aware that when you perform your search, if you do not find any results, it may be that the keyword you are using is not recognised in any of the fields describing the experiment. This does not always mean that your search has failed but simply that you might need to use a different keyword to perform your search.


Search results

Results from your refined search are now displayed (Figure 6). By reading the experiment titles you might identify a particularly interesting experiment and wish to find out more about it. All the information available for an experiment can be easily accessed by clicking on the accession number.

Figure 6 Search results




6. Click on the experiment accession to view more information about the experiment. In our example we selected experiment E-MEXP-862.




The experiment page

Let's find more detailed information on experiment E-MEXP-862 (Figure 7). The expanded view provides useful links to information such as data files and sample annotation.

Figure 7 Detailed information on experiment E-MEXP-862



[A] This section contains detailed sample information including sample characteristics and experimental factor values. The next page explains this section in more detail.

[B] In this section you can find details of the array used.

[C] In this section you can find details of the protocols used.

[D] This section contains the experiment desciption as supplied by the submitter

[E] Here you can find the keywords used to describe the experiment type. These keywords are added by the submitter.

[F] Here you can find the submitter contact details

[G] The MIAME score indicates what information required by the MIAME guidelines is available for the experiment. For HTS-based experiments the MINSEQE score is shown. A high MIAME or MINSEQE score indicates a well described experiment.

[H] This section contains the different MAGE-TAB Files describing the experiment: Investigation Description File (IDF), Sample and Data Relationship file (SDRF), Array Design file (ADF) and data archives.

[I] R objects, generated using version 2.11 of the ArrayExpress BioConductor package, are available for some experiments. These can be easily loaded into Bioconductor for downstream analysis.

[J] The ‘Browse all available files’ link takes you to a page listing all the files available to download for a particular experiment. More information on these files can be found in the section 'How to get data from the Archive'.

[K] This section might include links to:

  • Expression Atlas showing the most differentially expressed genes in the experiment

  • GenomeSpace ( is a data analysis environment that allows you to move data between different tools in order to analyze them. To use it, you need to get a GenomeSpace account by registering on their website. Then, when you find a dataset of interest in ArrayExpress, click on the link, log into genome space, select what you want to import (at the moment we transfer IDF, SDRF and data archives by default) and this will be loaded into GenomeSpace. From there you can select a tool you want to use for the analysis (GenePattern, Cytoscape, Genomica, IGV, etc) and the ArrayExpress files uploaded will be converted into the format required by the chosen tool.

Samples View

Let's continue exploring the sample information (A in Figure 7) for experiment E-MEXP-862.

This view (Figure 8) provides information on sample characteristics and experimental factors that are fundamental to understanding the results obtained in the experiment. Generally, each row corresponds to a sample and columns include sample characteristics and their relationship to the resulting data files. This provides a quick view over the structure of the experiment and the biological questions that the submitters addressed.

Figure 8 Detailed sample information and links to data



Notes[A] This column will provide you with the name of the experiment samples as provided by the submitter (this can be a number or text).

[B] In these columns you wil find important characteristics of the sample, such as the organism or cell type, as specified by the submitter. Sample characteristics will differ between experiments.

[C] In these columns you wil find the factor values. An experimental factor is a property which varies between samples and is important in the interpretation of an experiment, e.g. time, compound, genotype.

[D] In this column you will find links to the raw and processed data files.