This is going to seem odd, in fact, it may seem very odd.

However, you have already witnessed this problem in your earlier work on this site.

As you may have noticed, not all datasets (databases) use the same unique identification codes for the same thing. For example, you have been using a lot of Affymetrix IDs such as 218692_at, and you may have spotted that in some of the analyses you have been doing that these Affymetrix IDs map two different IDs in different systems. In fact, if we take 218692_at we will find that it maps to NP_001093226, NP_001093213, NP_001093223, NP_001093217, NP_001093219, NP_001093224, NP_001093221, NP_001093225, NP_001093222, NP_001093218, NP_001093216, NP_001093214 and NP_001093220 in the protein database at NCBI.

In an ideal world, all the different systems with DNA and protein sequences would easily map to the same unique ID. However, sadly this is not the case.

Sometimes in bioinformatics, we have to take IDs from one system and convert them to another so that we can continue our analysis. For example, we may have needed to map the protein ID - NP_001093226 - back to its Affymetrix ID - 218692_at - so we could carry out some analysis. Luckily for us, we don't have to do this by hand as there are a number of systems that can help us.

In this section, we are going to look at one such system which is the Ensembl implementation of a system called Biomart. Yes, the system looks scary but we will walk through it over the rest of the page. It's not that bad. Honest!

You have been carrying out a series of proteomics experiments that have produced a list of proteins that have changed expression levels, when compared to healthy patients, in the pancreas islets cells.

We are now going to look at how we can convert a list of protein IDs (we have used the RefSeq Protein ID (e.g. NP_001005353) identification system) to their equivalent Affymetrix IDs (Affy HG U133A probeset).

Setting up Biomart takes a bit of work. However, it is fairly straightforward for the type of conversion we are going to do. In fact, the system just needs four pieces of information:

We are now going to set up the system...

At the start the system looks like this:

The screen should now look like this:

Next, we want to set up the 'Filters', this is nothing complicated it's just really asking for the data you wish to convert.

Your screen should now look like this:

The completed setup should look like the image below.

We have now completed the settings to upload our data. We are now going to set the 'output' (also known as the 'Attributes').

The 'Attributes' section should now only have one item selected, that is, 'Affy HG U133A probeset'. It should look like the image below:

You have now set up Ensembl Biomart to convert a series of RefSeq Protein ID to there are Affymetrix equivalents.

If everything has worked okay then your screen should now look like the image below:

One last thing to do, and that is to download the data.

You should now have downloaded a file, the contents of which look a bit like the screenshot below:

We are now going to use this dataset with DAVID.

You have now successfully used DAVID to analyse two different datasets. You also learned how to successfully convert protein IDs to Affymetrix IDs.

The next stop is to use a system called String to find and map relationships between proteins.