Rational High-Throughput Screening of Cocrystals: An Explainable Deep Learning Framework Based on Morgan Fingerprints and Cross-Attention

Yukun Liu, Yanfei Liu, Ziang Du, Haochun Lan, Hao Zheng, Dongdong Yang, Yang Gao · 2026

Read Rational High-Throughput Screening of Cocrystals: An Explainable Deep Learning Framework Based on Morgan Fingerprints and Cross-Attention by Yukun Liu, Yanfei Liu, Ziang Du, Haochun Lan, Hao Zheng, Dongdong Yang, Yang Gao (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/PMC13614126.txt.

Description

Yukun Liu and colleagues present a 2026 deep learning system for predicting which pairs of molecules will form cocrystals, solid materials built from two compounds in one crystal. Drug makers use cocrystals to improve how medicines dissolve and last, yet hunting for good pairs by trial and error is slow and often blind. Earlier computer models gave weak detail and could not explain their own answers.

The new system reads molecules as Morgan fingerprints, a numeric sketch of their structure, and uses a cross-attention method to weigh how the two partners interact. It reached an accuracy score of 0.9878 on the main test measure. Extra checks compared rival models, varied settings, and used SHAP analysis to show which features drove each call. Chemists get both a fast screen and a look at the reasoning behind it.

Catalog record

CREATORYukun Liu, Yanfei Liu, Ziang Du, Haochun Lan, Hao Zheng, Dongdong Yang, Yang Gao
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE74902
LICENSEhttps://creativecommons.org/licenses/by/4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/article/MED/PMC13614126

Browse the full library · About the Tower of Babel