4 ms·
Apologies for a dumb question from someone not in the field. A protein's structure is also dependent on its environment, correct? Things like pH and temperature
by hn_throwaway_99 5y ago
Apologies for a dumb question from someone not in the field. A protein's structure is also dependent on its environment, correct? Things like pH and temperature, but also the fact that proteins bind to particular ions or other molecules. How do these factors figure into the "protein folding" problem?
- COGlory 5y agoThose factors do figure in, and AlphaFold has no way to address them, mainly because it's more a pattern matching algorithm than a protein fold computer. Experimental methods are also quite poor at addressing them.
- Gatsky 5y agoCorrect. Having said that most proteins we are interested in fold in a narrow range of pH and temperatures as maintained in a cell in a large organism.
- ehsankia 5y agoWouldn't that just be two more parameters in the training network?
- COGlory 5y agoNo, because there's little to no training data, and the neural network doesn't appear to actually have any real information on protein folding physics, beyond just very advanced template matching.
- up6w6 5y agoA related question I always had: how much does computational tools really help in the biochemistry as a whole given the tons of variables that we just can't even measure perfectly ? As a computer scientist I can see how rigid data structure algorithms are useful in dna sequencing but I'm not aware of how useful is a simulation of some metabolic pathway for example.
- pfisherman 5y agoComputational tools are insanely helpful and wifey used in biotech. Generally you see them spring up around experimental platforms (ex gene / RNA sequencing). In terms of your particular example, how good is your simulation? If your simulation can accurately predict the effects of perturbations on metabolic pathways, then PM me and we can start a biopharma company.
- up6w6 5y agoIt's exactly what I'm asking, what is the status of our current metabolic pathway simulations for example ? how much about protein denaturation do we need to know before protein structure prediction is useful ?
- pfisherman 5y ago> how much about protein denaturation do we need to know before protein structure prediction is useful ? Denaturation is not all that significant. Physiological conditions (i.e. pH, temperature, etc) tend to be maintained by homeostasis. Protein structure prediction is already incredibly useful. > what is the status of our current metabolic pathway simulations for example ? IMO we do not have the data to build useful metabolic pathway simulations. To do a good simulation you need to know (1) the proteins and metabolites involved, (2) their functions and key biophysical properties (ex. Km), (3) feedback / control elements. For most pathways we do not have (1). The data for (2) are mostly nonexistent, buried in literature, and / or unreliable. Same with (3).
- pfisherman 5y agoCorrect. Protein structures often adopt multiple conformations influenced by the factors you listed. There are things you can do - like molecular dynamics simulations - to study / predict protein conformations in silico. But you can’t do that without a structure.
- Rochus 5y agoIt's even much more complex. We e.g. know that many proteins get their "working conformation" only with help of other proteins (not only by the entropy guided physical processes), or that the conformation is flexible (i.e. the measured crystal structure is not representative).