CSC8312 -- Bioinformatics Theory and Applications

Introduction

One of the hopes of bioinformatics is to be able to use protein structure information to predict the function and behaviour of that protein. While we are some way of this at the moment, there are a large number of resources available for interacting with 3D protein structures.

In this practical, we are going to look some of the resources which are available and view some structures. In this practical, we are going to investigate the Opsin protein because it has a well characterised structure. It's also characteristic of a very important group of proteins.

What is Opsin? What group of proteins is it a member of? Why are these proteins so well studied?

Secondary Structure Prediction

The secondary structure of a protein is relatively straight-forward to predict from the primary structure (that is the amino acid sequence). A quick google search will reveal quite a large number of tools for secondary structure prediction. If you follow these, quite a few of them are dead or have not been heavily maintained; in general, most biologists turn to similarity searching in the first instance. Secondary structure prediction is not as useful as it once was.

Two of the better options are JPred or PSIPRED. There is also Predict Protein; this has one main advantage — a good name — and one main problem — a registration requirement.

Try either JPred or PSIPRED to make a secondary structure prediction for Uniprot protein OSPD_BOVINE.

Both of these take some time to run¹, so while they are doing so try the following questions.

How do these tools work? How many different techniques can you find for second structure prediction?

Homology Searching and PDB

If you used JPred, then it will have noted that your protein sequence had a highly similar sequence in PDB and suggested that you try PDB instead. This is good advice; let's go to the PDB Record.

Near the top is a icon that looks like some text; click on this. You should see the flat-file PDB record.

Like Uniprot, you should notice that there is a REST-style URL. Rather unusually, it's NOT a two-letter prefix flat-file. XML is also available.

Now, we can look at the sections of the PDB record.

Try the "Geometry" tab

This is pretty dull in this form; we will look at this in a more graphical form later; it's useful to realise, though, that the raw data is nothing mysterious. It's just some co-ordinates.

Moving on to the Sequence Details tab.

Here, you should see the overall secondary structure of the protein.

Do these correlate with the secondary structure predictions that you made earlier?

Investigate the "Materials and Methods" and "Biology and Chemistry" sections for yourself.

Finally, we consider the Structure Summary. On this page is the main thing that people think off when considering 3D structure. On the right, you should be able to get to the JMol viewer (feel free to try any of the others — you can also download RasMol and use that if you wish).

How does the protein structure that you are seeing relate to the function of the OPSD_BOVIN protein?

1. They can take so long, that I have put a search for JPred here.