Michelanglo — VENUS

A VENUS page is created anew on request. A Michelanglo page is stored and can be edited.

To make a page that best suits your needs please consider the following:

• The initial view will be the current view. Please choose the desired model and orientation (close this modal).
• The feature viewer will be in the modal that appears when the button at the bottom of the description is pressed.
• The blocks will be present in the description. Please rearrange and/or delete to suit (close this modal).

The following models can be included (please note about 20% of users visiting a page are on a mobile device and each structure is several megabytes).

Phosphorylation sites perform a variety of tasks. The next step would be to read if anything is known about this site. The Uniprot entry may provide some information. Some are known only from high throughput studies. Venus uses data from PhosphositePlus so a next step could be to see how common these are. It may be helpful to also check whether anything is known for paralogues of the protein. Also, if your variant is near a phosphorylated residue, but not one, check the "motif" section to see if a linear motif is predicted and whether it is likely to be affected (this is inexact, but can help).
Loss of a post-translational modification could mean there is a loss of regulation. For example:

• If your variant is a de novo heterozygous mutation, you are likely looking at a gain of function, wherein the phosphorylation was meant to keep the protein in check.
• If your variant is a compound heterozygous or homozygous, you are likely looking at a loss of function, wherein the phosphorylation was meant to activate the protein or allow an interaction with another protein.

Phosphorylation sites perform a variety of tasks. The next step would be to read if anything is known about this site. The Uniprot entry may provide some information. Some are known only from high throughput studies. Venus uses data from PhosphositePlus so a next step could be to see how common these are. It may be helpful to also check whether anything is known for paralogues of the protein. Also, if your variant is near a phosphorylated residue, but not one, check the "motif" section to see if a linear motif is predicted and whether it is likely to be affected (this is inexact, but can help).
Loss of a post-translational modification could mean there is a loss of regulation. For example:

• If your variant is a de novo heterozygous mutation, you are likely looking at a gain of function, wherein the phosphorylation was meant to keep the protein in check.
• If your variant is a compound heterozygous or homozygous, you are likely looking at a loss of function, wherein the phosphorylation was meant to activate the protein or allow an interaction with another protein.

Phosphorylation sites perform a variety of tasks. The next step would be to read if anything is known about this site. The Uniprot entry may provide some information. Some are known only from high throughput studies. Venus uses data from PhosphositePlus so a next step could be to see how common these are. It may be helpful to also check whether anything is known for paralogues of the protein. Also, if your variant is near a phosphorylated residue, but not one, check the "motif" section to see if a linear motif is predicted and whether it is likely to be affected (this is inexact, but can help).
Loss of a post-translational modification could mean there is a loss of regulation. For example:

• If your variant is a de novo heterozygous mutation, you are likely looking at a gain of function, wherein the phosphorylation was meant to keep the protein in check.
• If your variant is a compound heterozygous or homozygous, you are likely looking at a loss of function, wherein the phosphorylation was meant to activate the protein or allow an interaction with another protein.

Phosphorylation sites perform a variety of tasks. The next step would be to read if anything is known about this site. The Uniprot entry may provide some information. Some are known only from high throughput studies. Venus uses data from PhosphositePlus so a next step could be to see how common these are. It may be helpful to also check whether anything is known for paralogues of the protein. Also, if your variant is near a phosphorylated residue, but not one, check the "motif" section to see if a linear motif is predicted and whether it is likely to be affected (this is inexact, but can help).
Loss of a post-translational modification could mean there is a loss of regulation. For example:

• If your variant is a de novo heterozygous mutation, you are likely looking at a gain of function, wherein the phosphorylation was meant to keep the protein in check.
• If your variant is a compound heterozygous or homozygous, you are likely looking at a loss of function, wherein the phosphorylation was meant to activate the protein or allow an interaction with another protein.

Ubiquitination sites allow the protein's concentrations to be lowered. Venus uses data from PhosphositePlus so a next step could be to see how common these are. If this site is indeed common and your variant is a de novo heterozygous mutation, it is likely that your protein level are off balance. For example, if your protein is part of a complex where another protein is keeping it in balance (e.g. alpha subunit and beta-subunit of a G-protein), this would result in spurious 'activation'-like behaviour. The Uniprot entry may provide some information if this is the case.

Disulfides allow the protein to maintain structural cohesion in harsh environments. Disulfides are found in ER, Golgi or secreted protein in eukaryotes, and in the periplasmic or secreted protein in bacteria. A potential lack of calculated destabilisation does not indicate that the protein is stable outside of the cytoplasm. NB. that unfortunately the NGL.js library (used for the protein visualisation) does not show disulfides and cross-linked molecules connected if connected via LINK or SSBOND entries, only CONECT, which is non-standard. As a result the cross-link is present the calculations by PyRosetta, but visually it is not shown.

The mutation is predicted to lower the Gibbs free energy of folding according to force field calculations. A negative value is stabilising, and a value of 1-2 kcal/mol is neutral. A hydrogen bond has about 1-2 kcal/mol. For more see documentation.
A destabilised protein is likely to still partially fold, but may have exposed hydrophobic surfaces, and as a result is targeted for destruction by proteasomes. There is not a universal formula to convert ∆∆G to % degradation. These correlate within a protein, but larger domains are more tollerant to do disruption than smaller ones. Also, different taxonomic clades have different threshhold of thermostability —E. coli protein are more stable than human proteins. Export machinery is less tolerant to partially unfolded protein. But generally a ∆∆G over +5 kcal/mol is unlikely to be anything but deleterious.

Conservation data is taken from . ConsurfDB is a site that provides pre-calculated conservation profiles for PDB structures as calculated by the tool rates4sites. In the case of Swiss-Model structures, Venus migrates the residue grades (and the conservation colour) from the template that was used for threading.
Buried residues are expected to be more conserved in light of the fact that a mutation is generally structurally deleterious and may require a compensating mutation (i.e. epistatis ), however, surface mutations may also be conserved if they have a role, such as an interface with another protein.
At present, there is no conservation data provided for AlphaFold2 models, but will be implemented soon.
Reference Adi Ben Chorin, Gal Masrati, Amit Kessel, Aya Narunsky, Josef Sprinzak, Shlomtzion Lahav, Haim Ashkenazy, Nir Ben-Tal ConSurf-DB: An accessible repository for the evolutionary conservation patterns of the majority of PDB proteins Protein Science, 29:258–267
See also here for all relevant papers and resources

The distance between residues stated herein is the closest distance between any atom of the two residues in the given structure (before energy minimisation). Namely, it does not take into account

• whether the two residue's sidechains may rotate making them closer (for which an MD simulation would be required)
• whether the side chains moved during minimisation (of the wild type or mutant)
• whether the side chain was placed incorrectly by Coot

This column collates information from a few sources.

• Uniprot

  is a database of manually curated entries with pieces of information taken from the literature, which in turn will have more information on that residues making it a logical next step in the exploration of the possible effects of a variant. Reference Bateman, A., Martin, M. J., Orchard, S., Magrane, M., Agivetova, R., Ahmad, S., Alpi, E., Bowler-Barnett, E. H., Britto, R., Bursteinas, B., Bye-A-Jee, H., Coetzee, R., Cukura, A., Silva, A. Da, Denny, P., Dogan, T., Ebenezer, T. G., Fan, J., Castro, L. G., … Zhang, J. (2021). UniProt: The universal protein knowledgebase in 2021. Nucleic Acids Research, 49(D1), D480–D489.

• gnomAD

  (human substitions only)
  is a database that collects the variants found in the healthy wild type human population. As discussed in the hypothesis generation notes it is useful to see nearby residues for gnomAD variants as it is likely that they may have a similar effect to the substitution investigated. However, some mutations may be conservative, hence why the ∆∆G is calculated —greater than +2 kcal/mol is destabilising (to that conformation). Do note that if a variant is found in one or two individuals (as marked with ) it should be treated with caution. Also note that gnomAD does not cover the entirety of the healthy human population, therefore a back-of-the-envelope calculation is in order to say what are the chances of one or more individuals with two copies of that allele (homozygous). Reference Karczewski, K. J., Francioli, L. C., Tiao, G., Cummings, B. B., Alföldi, J., Wang, Q., Collins, R. L., Laricchia, K. M., Ganna, A., Birnbaum, D. P., Gauthier, L. D., Brand, H., Solomonson, M., Watts, N. A., Rhodes, D., Singer-Berk, M., England, E. M., Seaby, E. G., Kosmicki, J. A., … MacArthur, D. G. (2020). The mutational constraint spectrum quantified from variation in 141,456 humans. Nature, 581(7809), 434–443.

• ClinVar

  (human substitions only)
  collates indentified pathogenic variants. Reference Landrum, M. J., Chitipiralla, S., Brown, G. R., Chen, C., Gu, B., Hart, J., Hoffman, D., Jang, W., Kaur, K., Liu, C., Lyoshin, V., Maddipatla, Z., Maiti, R., Mitchell, J., O'Leary, N., Riley, G. R., Shi, W., Zhou, G., Schneider, V., Maglott, D., Holmes, J.B., Kattman, B. L. ClinVar: improvements to accessing data. Nucleic Acids Res. 2020;48(D1):D835-D844.

• PhosphoSitePlus

  (human substitions only)
  is a database collecting post-translational modifications from high throughput screens. This is useful as most post-translational modifications have not been investigated and some of which may have a strong functional role. Reference Guex, N., Peitsch, M. C., & Schwede, T. (2009). Automated comparative protein structure modeling with SWISS-MODEL and Swiss-PdbViewer: A historical perspective. ELECTROPHORESIS, 30(S1), S162–S173.

The mutation is in a transmembrane region as annotated in Uniprot. Mutations can have a variety of possible effects.

Note Venus calculates ∆∆G in a aqueous implicit solvent model even if the protein is marked as membrane bound, because it is a non-trivial process to set up due many corner cases. As a result do treat the difference in Gibbs free energy with caution.

A substitution to a charged residue in the membrane may result in an inability to be embedded —resulting in degradation by the quality control machinery, making it a loss of function variant.

In some protein, the transmembrane region alters conformation upon a conformational shift in an aqueous domains. In some protein transmembrane region features a pore or channel allowing ion or small molecule passage. This can be spotted by looking at the surface (show). In many cases, the channel is closed but opens upon a conformational shift. Hydrophobic residues that form a barrier are normally referred to as a "gate", while charged residues controlling specificity as a "filter". the analysis of a substitution affecting a conformational change requires at least two or more conformations representing the open and closed states. Such a variant may favour one over the other —closed → LoF, open → GoF. Do note the effect may be partial and a full loss of gating or filter may be lethal for the cell —akin to an ionophore.

All variants in ClinVar along with their neighbours were scores with the backend of Venus (see repo ). ClinVar does not frequently report the mode of inheritance (homozygous or heterozygous/de novo), therefore the data is both for homozygous and heterozygous. The split presented here using gnomADs pLI scores is to counter this. The impact in ClinVar can be unknown, conflicting, likely benign, benign, pathogenic, likely pathogenic and a dozen others. For the data presented here only the 'benign' and 'pathogenic' are presented. (Full dataset summary on GitHub). The summary of each variant in Venus is done on the frontend (JS), while the summary in the analysis was on a cluster (Python) therefore the categories differ.

Venus works at the protein level. It does not do operations at the nucleotide level. It works with Uniprot IDs not ENSEMBL ids, therefore these need to be converted.
Venus uses the Uniprot canonical sequences only —protein that differ from the Uniprot canonical protein will be converted to the mutation in the Uniprot canonical protein. The Uniprot canonical protein is in some instances different than the NCBI or, very rarely, the ENSEMBL one. If you have a mutation in a different format or not in humans, visit a conversion site, such Mutalyzer or VEP online .
To convert a genes (ENSG), transcripts (ENST) or protein (ENSP) (only one required) to Uniprot use the following: