Premium
Predicting Protein Glycation Rate and State: The Need for Models to Incorporate Additional Features
Author(s) -
Turkette Thomas,
Rhinesmith Tyler,
RootBernstein Robert
Publication year - 2019
Publication title -
the faseb journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.709
H-Index - 277
eISSN - 1530-6860
pISSN - 0892-6638
DOI - 10.1096/fasebj.2019.33.1_supplement.755.6
Subject(s) - glycation , chemistry , lysine , biochemistry , false positive paradox , amino acid , receptor , machine learning , computer science
Glucotoxicity is considered to be one of the primary drivers in type 2 diabetic (T2D) pathophysiology. As such, understanding the mechanisms underlying glucotoxicity is imperative in the identification, treatment, and prevention of T2D. One contributor to glucotoxicity is protein glycation in which a reducing sugar spontaneously reacts with the ɛ‐amino group of lysine residues or an N‐terminal amino group to form a covalently bound adduct. Here we test two methods employed to predict whether a given protein sequence will undergo glycation and whether either method offers insight toward predicting the rate of glycation in vitro . The first method involves looking for sequences that are homologous to known glucose binding regions, while the second utilizes NetGlycate‐ a neural network trained to predict glycation based on a protein sequence. Both methods were used to predict potential glycation sites on the human insulin receptor and regions containing these sites were synthesized. These fragments were then incubated in high glucose for 1, 3, or 6 days. The extent of glycation was assessed using MALDI‐ToF mass spectrometry. There was some overlap in the residues predicted to undergo glycation, but both methods also predicted unique sites. Both methods generated a number of false positives, but sites predicted by both methods were generally more likely to undergo glycation. Additionally, several of the sites predicted that were unique to one method underwent glycation. The rate of glycation for a given site varies greatly regardless of the method employed to predict it and this holds true even when both methods predicted the same site. Our results suggest: 1) neither method is sufficient to independently predict protein glycation consistently and accurately, 2) a model that incorporates both evolutionary considerations in addition to protein sequence would be expected to better predict protein glycation, and 3) neither approach informs the rate of protein glycation, suggesting that some other factor(s) must also be at play. This abstract is from the Experimental Biology 2019 Meeting. There is no full text article associated with this abstract published in The FASEB Journal .