Predicting polymerization reactions via transfer learning using chemical language models
Authors:
Brenda S. Ferrari,
Matteo Manica,
Ronaldo Giro,
Teodoro Laino,
Mathias B. Steiner
Abstract:
Polymers are candidate materials for a wide range of sustainability applications such as carbon capture and energy storage. However, computational polymer discovery lacks automated analysis of reaction pathways and stability assessment through retro-synthesis. Here, we report the first extension of transformer-based language models to polymerization reactions for both forward and retrosynthesis ta…
▽ More
Polymers are candidate materials for a wide range of sustainability applications such as carbon capture and energy storage. However, computational polymer discovery lacks automated analysis of reaction pathways and stability assessment through retro-synthesis. Here, we report the first extension of transformer-based language models to polymerization reactions for both forward and retrosynthesis tasks. To that end, we have curated a polymerization dataset for vinyl polymers covering reactions and retrosynthesis for representative homo-polymers and co-polymers. Overall, we obtain a forward model Top-4 accuracy of 80% and a backward model Top-4 accuracy of 60%. We further analyze the model performance with representative polymerization and retro-synthesis examples and evaluate its prediction quality from a materials science perspective.
△ Less
Submitted 17 October, 2023;
originally announced October 2023.
Estimating SARS-CoV-2 transmission in educational settings: a retrospective cohort study
Authors:
Mattia Manica,
Piero Poletti,
Silvia Deandrea,
Giansanto Mosconi,
Cinzia Ancarani,
Silvia Lodola,
Giorgio Guzzetta,
Valeria d'Andrea,
Valentina Marziano,
Agnese Zardini,
Filippo Trentini,
Anna Odone,
Marcello Tirani,
Marco Ajelli,
Stefano Merler
Abstract:
Background School closures and distance learning have been extensively applied to control SARS-CoV-2 transmission. Despite evidence of viral circulation in schools, the contribution of students and of in-person schooling to the transmission remains poorly quantified. Methods We analyze 976 exposure events, involving 460 positive individuals, as identified in early 2021 by routine surveillance and…
▽ More
Background School closures and distance learning have been extensively applied to control SARS-CoV-2 transmission. Despite evidence of viral circulation in schools, the contribution of students and of in-person schooling to the transmission remains poorly quantified. Methods We analyze 976 exposure events, involving 460 positive individuals, as identified in early 2021 by routine surveillance and through an extensive screening conducted on students, school personnel, and their household members during an outbreak in a small municipality of Italy. Results From the analysis of potential transmission chains, we estimated that, on average, 55.1%, 17.3% and 27.6% infection episodes were linked to household, school, and community contacts, respectively. Clusters originated from students or school personnel showed a larger average cluster size (3.32 vs 1.15), a larger average number of generations in the transmission chain (1.56 vs 1.17) and a larger set of associated close contacts (11.3 vs 3.15, on average). We found substantial transmission heterogeneities, with 20% positive individuals seeding 75-80 of all transmission. A higher proportion of infected individuals causing onward transmission was found among students (48.8% vs 29.9%, on average), who also caused a markedly higher number of secondary cases (mean: 1.3 vs 0.5). Conclusions Uncontrolled transmission at school could disrupt the regular conduct of teaching activities, likely seeding the transmission into other settings, and increasing the burden on contact-tracing operations.
△ Less
Submitted 28 March, 2022;
originally announced March 2022.
