3 ms·
> Most genetic disorders involve a missequenced protein. Most known do, but thats just due to identification bias. We do not even know how many genetic disorde
by patall 9y ago
> Most genetic disorders involve a missequenced protein.
Most known do, but thats just due to identification bias. We do not even know how many genetic disorders there are. Nor how many people are affected. Nor can we identify most of those that are of complex origin or due to larger genomic alterations / structual variations.
- yorwba 9y agoCan you give some examples of genetic disorders that do not involve a missequenced protein? I can think of missing/duplicated chromosomes, but those should be easier to detect than smaller differences.
- hirenj 9y agoLots more places things can go wrong not only before protein is synthesised, but also after too. An example is anything that hits a regulatory region for a protein (i.e. outside protein coding region). Chances are phenotype may not be as strong because you may end up with only partial transcriptional loss. I'm guessing it's super critical where you have developmental diseases, and you hit a TF binding site specific for development.
- patall 9y agoThere are many things that can go wrong with a protein. The simplest is misregulation, like a regulatoy motif being disrupted, gene architecture (TAD) being changed, splicing changes (somehow linked to missequence proteins but only via changed ration of different splice isoforms) or mRNA being degraded faster or slower (UTR mutations), all leading to different protein concentrations (if its a protein being affected, might as well be some RNA molecule). There is not that much known about these diseases so its hard to give typical examples. Here is an example of a tool that tries to predict such variants via ML, they curated a few hundered variants via literature search: https://www.ncbi.nlm.nih.gov/pubmed/27569544 https://www.ncbi.nlm.nih.gov/pubmed/27569544