In HumanMine, we load several datasets that provide information about expression data in different tissues. These datasets use different experimental techniques and measurements, making it difficult to compare them directly, but we have specific queries and visualizations for each dataset. The table below summarises these datasets.
|Data Set||Source||Measurement type|
|Affymetrix array||Array express E-MTAB-62||T-statistic, P-value, UP, DOWN, notDE|
|RNA-seq||Array Express Illumina body map E-MTAB-513||FPKM|
|Protein localisation (Ab Staining)||Protein Atlas||High, Medium, Low, Not detected.|
To compare expressions between different datasets, it is easier to make use of the list operations. For example, if you want to find all genes expressed in the brain according to all the datasets, you need to:
- query each set independently and create a list of genes at each step.
- Use the list intersection operation to create a stringent, high confidence set.
- Or use the list union operation to create a set of lower stringency, which contains a gene even if it only shows expression in tissue x in one dataset.
It is possible to identify genes expressed solely in one tissue (i.e. tissue-specific expression). You will have to decide on your expression cut-off for datasets that provide only TPM or FPKM values. Simultaneously, for the E-MTAB-62 and Protein atlas localisation data, it is possible to use the cut-offs already applied to the data by filtering for UP (E-MTAB-62) or High and Medium (protein atlas localisation).
The easiest way to view the available expression data for a single gene is to look at the Report Pages for that gene. Gene report pages display graphs showing the expression of a particular gene in human tissues. For a list of genes, similar graphs are available on the List analysis pages.
Many template searches are available to analyse the various sources of expression data. These templates provide details of the specific dataset. However, due to the widely variant nature of the different expression sources, it is generally necessary to query any dataset independently to identify genes expressed in a tissue - or set of tissues - or identify the tissues that a particular gene or list of genes are expressed in.
To view all template searches for expression data, you need to navigate to the Templates tab and then filter the templates list by “Expression” using the “Filter by category” subtabs or the categories tags.
This is a three step process:
Find and save the genes expressed in all tissues except the one you are interested in.
- To accomplish this, you can use the Tissue → Protein Atlas Expression template, which returns a list of genes that are localised in a given human tissue. Use the drop-down list to select all tissues one at a time except the one you want to find the tissue-specific genes, for example the Adipose tissue. The operator should be set to “One of” as shown in the screenshot.
- The template also offers two other filters, the expression level and experimental confidence/reliability. First remove the “not detected” and “low” options from the level constraint.
- Then remove the “uncertain” option from the reliability constraint.
- You can now create a list of all the genes returned by your search, which should appear at the top of the Lists view page. If you are unsure how to save a list of genes, see Save a result set for further analysis.
Find the genes expressed in the tissue you are interested in
- You can use the same template as above - just change the tissue selection to the one you are interested in.
- Again, save the set of genes returned by your search
Find the genes expressed ONLY in the tissue you are interested in. To find the genes that are expressed ONLY in the tissue you are interested in you need to subtract the first list created above from the second list.
To do this, navigate to your saved lists under the Lists tab;
new lists appear at the top of the Lists view page.
Select the two lists and then use the “Subtract lists” operation, which provides options to subtract the lists either way. You need to select the single tissue list - the “Adipose tissue” list - minus the all “tissues but one” list.
Enter a name for your new list, which should appear at the top of the Lists view page.