/sci/ - Science & Math » Thread #13423988

279KiB, 1228x1036, E66b7TvWQAcvAKT.jpg

View Same Google iqdb SauceNAO

biopunk 2037 incoming

Anonymous Thu 22 Jul 17:49:48 2021 No.13423988 View Reply Original Report

Quoted By: >>13424024 >>13424304

>Highly accurate protein structure prediction
for the human proteome
oh boy, can predict the shape of 98.5% of human proteins

Anonymous

View Same Google iqdb SauceNAO c.png, 447KiB, 774x699

Anonymous Thu 22 Jul 2021 18:01:37 No.13424024 Report

Quoted By:

>>13423988
>Despite this intensive study, only 35% of human proteins map to a PDB entry, and in many cases the structure covers just a fragment of the sequence

>The final dataset covers 98.5% of human proteins with a full-chain prediction.

>58% of residues with a confdent prediction, of which a subset (36% of all residues) have very
high confdence.

>Highlighted predictions:
>Glucose-6-phosphatase: a membrane-bound enzyme that catalyzes the final step in glucose synthesis; it is therefore of critical importance to maintaining blood sugar levels. No experimental structure exists,

>Diacylglycerol O-acyltransferase 2: for storing excess metabolic energy as fat in adipose tissue
>With our highly confident predicted structure (median pLDDT 95.9), we set out to identify the binding pocket for a known inhibitor, PF-06424439

>Wolframin: Mutations in the WFS1 gene are associated with Wolfram syndrome 1, a neurodegenerative disease
> A recent evolutionary analysis suggested domains for wolframin, which our prediction largely supports

Anonymous

Anonymous Thu 22 Jul 2021 19:25:33 No.13424304 Report

Quoted By: >>13424556

>>13423988
How do they calculate their confidence? From my experience biologists will always throw any computational predictions out the window. At the end of the day, X-ray crystallography will continue to reign supreme just because people are most familiar with it and they trust it.

Anonymous

Anonymous Thu 22 Jul 2021 20:33:22 No.13424556 Report

Quoted By:

>>13424304
this is true, models aren't seen as accurate because active sites can be as small as a hydrogen bond, which the 98.5% could certainly miss. models are more used for de novo synthesis of proteins, which are then tested experimentally after

Capcode	All Only User Posts Only Moderator Posts Only Admin Posts Only Developer Posts
Show Posts	All Only With Images Only Without Images
Deleted Posts	All Only Deleted Posts Only Non-Deleted Posts
Ghost Posts	All Only Ghost Posts Only Non-Ghost Posts
Post Type	All Only Sticky Threads Only Opening Posts Only Reply Posts
Results	All Grouped By Threads
Order	Latest Posts First Oldest Posts First

Your latest searches