Dataset for discovering new hypertension small molecules using machine learning-aided computational fragment-based...

Area of Science:

• Computational chemistry and cheminformatics
• Pharmacology and drug discovery
• Machine learning in medicinal chemistry

Background:

• Hypertension management often involves targeting the renin-angiotensin-aldosterone system (RAAS).
• Approved drugs include Angiotensin-Converting Enzyme Inhibitors (ACEIs) and Angiotensin II Receptor Blockers (ARBs).
• Novel drug discovery for hypertension requires efficient methods to generate diverse lead molecules.

Purpose of the Study:

• To demonstrate the utility of computational fragmentation-based and machine learning-aided drug discovery for generating novel antihypertensive agents.
• To create new lead molecules targeting the RAAS, specifically ACEIs and ARBs.
• To provide datasets that can accelerate the design of new drugs for hypertension.

Main Methods:

• A preliminary dataset of 63 molecular fragments from approved ACEI and ARB molecules was compiled from ChEMBL and DrugBank.
• New lead molecules were generated using computational fragmentation and machine learning approaches.
• Generated molecules were screened for oral drug criteria and ACEI/ARB functional group presence, then clustered using unsupervised machine learning.

Main Results:

• Three distinct datasets of newly generated molecules were produced: novel ACEIs, novel ARBs, and unassigned class molecules.
• The process successfully identified potential new lead compounds with desired pharmacological properties.
• The machine learning clustering effectively categorized molecules based on functional group allocation, aligning with drug classes.

Conclusions:

• Computational fragmentation and machine learning offer an efficient strategy for discovering novel antihypertensive drug leads.
• The generated datasets can significantly aid in the timely design of new antihypertensive medications.
• The developed model is adaptable for generating lead molecules for other therapeutic areas beyond hypertension.