<aside> 💵 5500 VITA ➕ 3500 USDC | Longevity Hackathon 1000 USDT | PrivateAI

</aside>

🟥 Introduction

Lifelong accumulation of advanced glycation end products (AGEs) contributes to the development of age-related pathologies, so that non-enzymatic modification of long-turnover proteins—like those in the extracellular matrix (ECM)—has been proposed as a new hallmark of aging (1). Consequently, the removal of AGEs and/or prevention of their formation is a promising therapeutic modality aimed at mitigating physiological deterioration associated with advanced age.

🟥 Presentation

Project pitchdeck | Psi Combinator

https://youtu.be/Gt-XDLevySw?feature=shared

🟥 Goal

Our goal is the creation of therapies aimed at AGE removal and/or prevention, using a ML/AI modeling toolkit, in silico screens, as well as the subsequent (pre-)clinical product development.

🟥 Scope

• Enzymatic or chemical breakage of covalent crosslinks formed as a result of non-enzymatic glycosylation (glycation) of the ECM proteins, such as collagens (including means of delivery, ADMET, etc.)
• Engineered biologics that remove noxious metabolic precursors of AGEs.
• Small molecule inhibitors/attenuators of glycation: dicarbonyl scavengers, transition metal chelators, transglycators, etc.
• Stimulators of cellular defense mechanisms against electrophilic stress.

🟥 Product development tracks

Tentatively, our approach to alleviating the glycation burden is conceptually organized into four parallel product development tracks:

[HACK-1]

De novo active site modeling based on substrate and catalyzed reaction inputs. Essentially, we’ll build an active site around the substrate of choice so that the desired reaction is catalyzed. From the practical standpoint, the most interesting substrate would be glucosepane, which is the most abundant crosslink in the ECM (2).

[HACK-2]

Predicted protein conformations screening for suitable active sites. Recently, Meta.AI reported on predicting conformations of 600M proteins from metagenomic sequence data (3). Our task would be screening these proteins for a suitable active site against modeled catalytic determinants from the previous step.

[HACK-3]