How to Find Expected Gene Expression Levels in Different Tissues for Use as qPCR Controls

Sometimes your qPCR results may exhibit very low levels of gene expression in your samples or controls. If this is unexpected, you may wonder whether your results are accurate, if the problem lies with the experiment's design or if your sample's integrity is compromised.

One option is to run the experiment again. However, there are online resources, such as the National Center for Biotechnology Information (NCBI) database, that enable you to check if gene expression results for a specific tissue type are in line with results from other researchers. Consulting the NCBI database is quick and easy, and helps you avoid a laborious literature search. Accessing database information could give you more confidence in your results, or it might help you to optimize your experiment or experimental design. This reference material could also help you choose an appropriate control tissue to validate your experiment. Read on for some examples of situations in which researchers find these resources useful, and discover how to use them with confidence.

Let’s look at some examples where researchers benefited from consulting an online database of gene expression information.

Scenario 1: Researchers investigated vascular endothelial growth factor A (VEGFA) gene expression in human lymph tissue cells treated with a drug known to increase VEGFA gene expression. Their qPCR experiment indicated negligible VEGFA expression in both the treated cells and untreated controls.

Was this negligible expression a genuine biological phenomenon, or was it caused by a problem with the qPCR experiment? To help solve this puzzle, the researchers used the NCBI database to find the expected VEGFA gene expression in human lymph tissue.

Result: According to the database information, VEGFA expression is considered to be extremely low or negligible in human lymph tissue, which means that the results of the researchers’ qPCR experiment were likely to be accurate. To confirm this, the researchers used the online database again to find a tissue expected to express VEGFA in higher quantities, and identified human thyroid tissue.

To validate their qPCR experiment, the researchers used cells from human thyroid tissue as a positive control. They observed significant VEGFA expression in the thyroid tissue sample, and again noted negligible expression in the lymph tissue samples. This confirmed to the researchers that their qPCR experiment was working as expected, and their results were accurate.

Scenario 2: Researchers investigated Ephrin B2 (Efnb2) gene expression in mouse kidney tissue following treatment with a pro-angiogenic drug. They performed a qPCR experiment using the treated kidney tissue samples and matching control samples, and observed negligible Efnb2 expression in the samples from both treated mice and untreated control mice. The researchers then consulted the NCBI database to check the expected levels of Efnb2 expression in mouse kidney tissue.

Result: Using the online database, the researchers found that mouse kidney tissue is expected to express Efnb2, but at a relatively low level compared with other mouse tissues. They used the database to find a mouse tissue with high levels of Efnb2 expression, and identified joint tissue. Using cells from mouse joint tissue as a positive control, the researchers confirmed that their qPCR experiment could detect Efnb2 expression. They suspected that the low Efnb2 levels in the kidney samples had made it difficult to detect.

The researchers optimized their qPCR experiment to better detect Efnb2 gene expression at low levels. They increased the amount of complementary DNA (cDNA) in the qPCR reactions and chose a different housekeeping gene that mouse kidney tissue typically expressed at low levels. Following these changes, the researchers were able to successfully detect Efnb2 expression in the mouse kidney samples, and identified a significant increase in Efnb2 expression in their drug-treated samples.

In the scenarios above, both research teams searched for expected levels of gene expression after their qPCR experiments had produced unexpected results. It might also be useful to consult the NCBI database before conducting your experiments, particularly if you are using a specific cell or tissue source, or investigating a specific gene for the first time. This would enable you to choose an appropriate control tissue and view expected gene expression levels for your chosen tissue in advance. This could help you to anticipate your results more accurately, or to preemptively optimize your experimental procedure for very high or low levels of gene expression.

With these two scenarios in mind, let’s learn how to use these online resources. Let's use Scenario 1, VEGFA gene expression in human lymph tissue, as a step-by-step example.

1. Visit the online NCBI Gene Database. This is a publicly available online database comprising gene-expression data, such as those from microarray experiments. These gene-expression data typically list the species and tissue source, and sometimes include data from diseased tissues, such as tumor samples.
   
   
2. Type the gene name, “VEGFA”, into the textbox and click 'Search'.
   
   
3. A list of gene records appears. The top hit is for the human gene, with the name “vascular endothelial growth factor A [Homo sapiens (human)].” The other records show the VEGFA gene in other species. Click the entry for the human gene.
   
   
4. You are now on the gene record page for VEGFA. If you scroll down, you will see plenty of information. Right now, we are interested in expected levels of expression in different tissues. To see relative expression levels in different tissues, click the 'Expression' link on the right side of the page in the 'Table of Contents' section. This will display the relevant part of the gene record page. For the VEGFA gene, you will see a graph showing relative gene expression data for different human tissues. The data for this gene record are based on high-throughput RNA sequencing for 27 different human tissues from 95 human volunteers. The graph is ideal for a quick first look at expected levels of gene expression. The dataset does not contain the specific tissue we are interested in ("lymph tissue"), but it does include “lymph node”, which shows very low expression compared with other tissues. Nevertheless, we will need to look elsewhere to find the “lymph tissue” we are interested in.
   
   
5. On the right side of the gene record, click the 'Unigene' link. This will link through to the VEGFA Unigene record. Click on the record to access it (it should be the only one listed), scroll down to the 'Gene Expression' section and click the 'EST Profile' link to access the VEGFA EST Profile. Expressed Sequence Tags (ESTs), are short nucleotide sequences of 200–500 bases and represent a small portion of a gene. Scientists can submit EST data to the NCBI database along with the name of the tissue in which they found the ESTs. The EST profile for a gene contains a tabular and graphical representation of relative EST numbers found in specific tissues for a given gene. This provides approximate relative expression levels.
   
   
6. The VEGFA EST Profile provides EST counts for various human tissues. The ratios beside each tissue represent EST numbers for that specific gene, divided by the number of ESTs in that group. The database also provides a normalized expression value, in transcripts per million, so you can easily compare expression levels between tissues. If you look for “lymph tissue” in the list on the VEGFA EST profile, you will see that there are no detected ESTs in the recorded lymph samples. If you scan the list to identify a control tissue that should express VEGFA in greater quantities, you will find several candidates, including thyroid tissue, which we expect to express VEGFA at 343 transcripts per million. A thyroid tissue sample is therefore a potentially suitable positive control for qPCR analysis of VEGFA expression.

Although using the Unigene EST database is useful and convenient, this database only approximates expected expression in specific tissues, and is not a highly accurate or comprehensive guide to gene expression levels. The tissue or cell samples that generate the cDNA libraries are not strictly defined, which means the database provides a largely crude measure of expected gene expression. Nevertheless, it is a useful first port of call when troubleshooting unexpected qPCR results, and can help identify potential control tissue samples and plan experiments.