At the starting page of each 3DM database, you see the 3DM data cycle. The icons in the circle represent links to the most important 3DM options. These options are also available on the left.

In the keyword search tab of the search option, you can select species. Here you can type the species names. We have separately searched for each of the species and the resulting proteins were added to the subset window by clicking the + signs of the subset window. Try to search yourself for Aspergillus clavatus in the species search options. You will find two proteins (A1CFP3 and A1CMM8). These are the first two proteins of the subset. You can get a list of the proteins that are in a subset by clicking in the subset window on the number (in this case 33) that indicates how many proteins are in the subset.

How many positions are 100% conserved? And how many of those are specific for the oxalate-producing fungi?

Difference = 98.94%. In the "subset specific conserved" plot the number is 99,38 (you can see this by putting your mouse over the peak of the subset specific conserved plot at position 157). The difference between these numbers come comes from the fact that in the full alignment the serine's of the oxalate producers subset are included. The "subset specific conserved residues" plot calculates the difference between the conservation in the subset minus the conservation of the full set but then without the sequences of the subset.

The proteins in a superfamily usually form groups of different specificities. Within a group the residues important for specificity are conserved, but between the groups, they mutate. Since they all mutate simultaneously between the groups they result in a correlated mutation network. Note: which protein feature is behind a correlated mutation network heavily depends on the input alignment. If you make a subset of enzymes that all have the same specificity the correlated mutation will, of course, not reflect changes in specificity and thus the network will not be composed of specificity hotspots. The concept of how to choose the input alignment is explained in more detail later in this practical.

Position 157 is the highest correlating positionsposition. Note that you can click on the heatmap. This will lead to plots showing the amino acid distribution of the two corresponding positions. Those plots show the co-occurrence of amino acids. This data can be use to see if it might be better to make certain double mutants instead of single mutants.

It enables to the plot of any other data type of which the 3D number is known on the network. This is one of the strong feature features of 3DM. All data and all tools are connected via the 3D numbers.

The enrichment score is the factor that shows how many more times mutations related to the keyword are found in the network compared to positions outside the network. The enrichment score for specificity in the OAH network is 7.39. This means that there are 7.39 times more mutations effecting affecting specificity published at positions in the network. Thus it is likely that specificity is the evolutionary pressure that caused the positions in the network to mutate simultaneously.

Note that an enrichment score of >4 or 5 normally is significant.

Position 157 scores highest when both numbers are added. Positions that both make a contact with a ligand but also show correlated mutation behaviour behavior are likely hotspots for specificity.

What is the residue type of 157? In yasara you can make a residue visible by right-clicking on the residue in the sequence at the bottom of Yasara and choose "Show Atoms → residue"

No, in a subset with only P on position 157 the P is, of course, 100% conserved and therefore can't mutate together with other positions. The correlated mutation data that 3DM calculates is a measure for of how often mutations occur together between two positions. No mutations will result in a score of 0.

In the figure, you can see where the correlated mutation are is found in the P157 alignment. This plot was generated by clicking "visualize all notes" in CorNet when the P157 subset was selected.

From within Yasara you can also make this visualization by the following functions:

• 3DM → select subset → P157
• 3DM → show superfamily data → Correlations

Now to answer the question do the following:

• 3DM → select dataset → full dataset
• 3DM → show superfamily data → Correlations

You can switch between the two scenes using the different tabs. Clearly, the two scenes overlap. So the correlated mutations in the P157 alignment are formed during evolution in the dimerization domain.

This question can be answered thoroughly or just very simple. One thing is for sure, everything shows that the correlated mutations are important for specificity. Those are your first-choice hotspots. Then pick as many positions as your screen allows only using common residues at these hotspots (e.g. residues > 2% or so). This cufcut-off percentage is depending on how many hotspots you want to use. The more hotpots the higher this percentage or your library size gets too big. So it is always a trade-off between the number of hotspots and the number of residues per hotspot. There are many things to consider when you choose hotspots and the residues at the hotspots. Each position should be considered carefully and all data at each position should be investigated.

These are things to consider:

• Read the articles that describe mutations at the hotspots. They might have a role that you don't want to target. The template sequence might also have things specific for the template that should not be touched.
• For the same reasons look in at the structure. Am I not destroying a salt-bridge unique to the template. If so, you probably need to mutate the other partner of the salt-bridge as well. Although I would probably not touch the bridge and choose another hotspot if this position is not THE hotspot.
• Look at which proteins have the residue that I want to introduce. So for each position you look for each residue what type of protein have has this amino acid. When choosing the residues to which you want to mutate you sometimes might want to exclude residues that are just above your cut-off. You might want to exclude hydrophilic residues, for instance, if those are only present in enzymes that do an exotic reaction. Try to downsize your library this way by excluding residues per position.
• Check all data types at each hotspot you want to target to see if it doesn't have a role (e.g. dimerization) that might harm the activity.
• Always use your brain to find the best combination of hotspots and amino acids. It is a tricky business and you will get better by practice.
• Make the library bigger than you can screen. If correctly designed your library will contain more than one hit and you don't need to find all hits. Even if it does contain just one hit, you also don't need to screen 98% of your library to find the hit. Screening 70% should give you enough confident confidence that if no improvement is detected you are better of designing a new library.