The overall goal is to produce a visualization tool, or visualization, that accurately suggests mutations that can be added to dTIM to rescue functionality, and that ideally aids the user in understanding why those mutations, in particular, were problematic for functionality, while others were innocuous.

We suggest that you consider some of the following when developing your visualization, though of course, these are mere suggestions. If you have better ideas, you are definitely encouraged to pursue them:

  • Agreement with the family consensus - that is, fitting within the residue distribution as defined by other functional members of the family - is almost certainly not a primary requisite for functionality. Our broken dTIM was created by trying to make a particular species variant of TIM fit the consensus better. As a result, comparisons of the parent (S. cerevisiae) to the family consensus, and some representation of the distance between dTIM and its parent, may be more productive than comparisons of dTIM to the family.
  • The functional deficit caused by some problematic mutations in dTIM can be reversed by mutating them back to their parent residue. Others can be reversed by changing _different_ residues that minimize the impact of the residue already present.
  • The extent to which each residue is constrained to a particular identity, or subset of biophysical properties, varies across the protein. This can be considered a sort of uncertainty. Casting this data into the three-dimensional structure may provide insight into what portions of the protein are generally allowed to change, and which are constrained. Because sequence proximity and spatial proximity are only loosely correlated, examining the spatial locations of mutations may provide insight into their likely acceptability in a fashion that cannot be seen by examining the sequence.
  • A mutation, because it changes a subset of the atoms of the protein, also changes the network of which parts of the protein touch what. This can modify, for example, energy transport mechanisms that require passing an electron from amino acid side-chain to side-chain. It can also alter protein flexibility, or the ensemble of conformations that the protein can adopt.
  • Instructions for running molecular dynamics simulations will be provided in the forum, if you would like to examine how the network of contacts might change in the presence or absence of specific mutations, or how mutations affect the way that the network changes when the protein flexes.
  • Any of the structural-biophysical-volume properties of the protein that can be painted onto a sequence or multiple-sequence-alignment view of the protein, will dramatically aid end-users in applying their domain knowledge and intuition to the problem.
  • Any of the multiple sequence alignment properties that can be painted onto the structure, likewise would be quite useful to the working bio/life-sciences researcher.
  • Not all proteins have structures determined for them. Sequence data is much more easily available, enabling more powerful statistical inferences.
  • Evolution has clearly voiced its opinion regarding requirements for functionality, some of the artifacts of which can be found by looking for groups of residues that co-evolve (that is, where one residue changes, others always change with it) within the family. Identifying these groups and displaying their membership on either the multiple sequence alignment, or on the protein structure, could be very informative.