Single sequence tutorial

The selection below enables you to pick the protein you want to display in the plot below. The protein identifiers are taken from the FASTA file that you submitted, and cannot be changed here. In example below, we have include proteins that form TIM barrels, in relation to an article describing A Comparative Thermostability Structural Analysis of Ten Different TIM Structures .

We also added the sTIM-11 protein , which was designed de novo by the Baker group and folds as a TIM barrel, and the OctaV1 protein , which was also designed to fold as a TIM barrel, but in fact folds differently, even though its secondary structure elements are generally in the intended (TIM barrel) positions. The multiple sequence alignment we are using was generated with the PROMALS3D server based on the 11 TIM barrel structures, with OctaV1 added based on the location of its secondary structure elements.

The plot below displays the predictions (y-axis) in relation to the protein sequence for the selected protein (x-axis). By default, the first 200 residues of the sequence are displayed. You can modify the range of residues displayed by either changing the numbers in the boxes below, or by changing the slider. The plot will be automatically rescaled if you do so, note that this can take some time for longer sequences.

When you run a specific predictor, for example DisoMine, only that data will initially be displayed, but all predictions that were executed to obtain the final one will be available - just click on their name below the plot to toggle individual predictions on and off. When you hover over the plot, it will display the amino acid residue that the cursor is currently on (x-axis), and the corresponding prediction values are listed next to their name at the bottom of the plot. A brief explanation of the interpretation for the predictions is available by 'Click for explanation'.

If you turn off the 'sidechain', 'coil', and 'ppII' values, you will notice that the proteins we included all have low disoMine (disorder) values and dynaMine backbone values of generally above 0.8, in agreement with the fact that these are well-folded proteins. For the TIM barrel proteins, there are consistent peaks for beta sheet propensities, which tend to coincide with the early folding peaks, with helix propensities generally of similar maximum height as the beta sheet.

Displaying the designed sTIM-11 protein evidences the subunit repeats, as groups of signals are repeated 4 times, with some differences because of the fine-tuning of their final design. This protein is shorter, but the overall pattern of the signals is similar to the natural TIM barrel proteins.

However, the OctaV1 protein, despite being designed based on a TIM barrel structure template, shows a quite different pattern, with much higher helix propensities in relation to beta sheet and early folding peaks corresponding to the regions with high helical propensity. This indicates that its 'emergent' behavior, or what the OctaV1 sequence is capable of doing based on local amino acid interactions, is quite different from TIM barrel proteins.

In other words, the local interactions between amino acids in the OctaV1 sequence seem to determine a different overall behavior, which might especially influence the secondary structures that are formed first when folding, and so change the fold it finally obtains. The MSA part of the tutorial better illustrates this from on a structure-based MSA of these proteins.