Dr. Rocío Mercado Oropeza

Dr. Rocío Mercado Oropeza

Assistant Professor

Sitemap

A list of all the posts and pages found on the site. For you robots out there is an XML version available for digesting as well.

Pages

publications

The materials genome in action: identifying the performance limits for methane storage

Published in Energy Environ. Sci., 2015

In this Perspective, we have collected and compared on a consistent basis the methane uptake in over 650 000 existing and predicted materials using molecular simulations. The results suggest that it may be difficult to reach the (at the time) Advanced Research Project Agency-Energy (ARPA-E) target for natural gas storage.

Recommended citation: Simon, Cory M. et al. (2015). "The materials genome in action: identifying the performance limits for methane storage." Energy Environ. Sci. 8. 1190-1199. link

What are the best materials to separate a xenon/krypton mixture?

Published in Chem. Mater., 2015

In this work, we screened the Nanoporous Materials Genome, a database of about 670 000 porous material structures, for candidate adsorbents for Xe/Kr separations using a hybrid computational approach combining random forests with molecular simulations. Our study predicts that the two most selective materials in the database are an aluminophosphate zeolite analogue and a calcium-based coordination network, both of which had already been synthesized but (at the time) not yet tested for Xe/Kr separations.

Recommended citation: Simon, Cory M. et al. (2015). "What are the best materials to separate a xenon/krypton mixture?" Chem. Mater. 27(12). 4459-4475. link

Computer-aided search for materials to store natural gas for vehicles

Published in Front. Young Minds., 2015

In this article (written for kids aged 13-15), we show how computers can be used to search for the most promising sponge-like materials for storing natural gas.

Recommended citation: Simon, Cory M. et al. (2015). "Computer-aided search for materials to store natural gas for vehicles." Front. Young Minds. 3-11. link

Systematic tuning and multifunctionalization of covalent organic polymers for enhanced carbon capture

Published in J. Am. Chem. Soc., 2015

Here, we synthesized 17 novel porous covalent organic polymers with a wide range of properties, achieved by tailoring the length and geometry of the building blocks. The multifunctionalization strategy presented can be used in the design of novel catalysts, the synthesis of functional sensors, and the improvement of existing covalent organic polymers.

Recommended citation: Xiang, Zhonghua et al. (2015). "Systematic tuning and multifunctionalization of covalent organic polymers for enhanced carbon capture." J. Am. Chem. Soc. 137(41). 13301-13307. link

Force field development from periodic density functional theory calculations for gas separation applications using metal–organic frameworks

Published in J. Phys. Chem. C, 2016

Here, we present accurate force fields developed from density functional theory (DFT) calculations with periodic boundary conditions for use in molecular simulations involving MOF-74 and related frameworks, where conventional force fields fail to accurately model gas adsorption due to the strongly binding open-metal sites. The introduced DFT-derived force fields predict the adsorption of CO2, H2O, and CH4 inside these frameworks much more accurately than other common force fields.

Recommended citation: Mercado, Rocío et al. (2016). "Force field development from periodic density functional theory calculations for gas separation applications using metal–organic frameworks." J. Phys. Chem. C. 120(23). 12590-12604. link

Unexpected diffusion anisotropy of carbon dioxide in the metal–organic framework Zn2(dobpdc)

Published in J. Am. Chem. Soc., 2018

Here, we investigate the diffusion of CO2 within the pores of Zn2(dobpdc) using pulsed field gradient nuclear magnetic resonance spectroscopy and molecular dynamics simulations. In addition to observing CO2 diffusion through the channels parallel to the crystallographic c-axis, we unexpectedly find that CO2 is also able to diffuse between the hexagonal channels in the crystallographic ab plane, despite the walls of these channels appearing impermeable in molecular dynamics simulations. This suggests the presence of defects that enable effective multidimensional CO2 transport in a metal–organic framework with nominally one-dimensional porosity.

Recommended citation: Forse, Alexander C. et al. (2018). "Unexpected diffusion anisotropy of carbon dioxide in the metal–organic framework Zn2(dobpdc)." J. Am. Chem. Soc. 140(5). 1663-1673. link

In silico design of 2D and 3D covalent organic frameworks for methane storage applications

Published in Chem. Mater., 2018

Here, we present a database of 69 840 covalent organic frameworks assembled in silico from 666 distinct organic linkers and four established synthetic routes. Due to their light weights and high internal surface areas, we investigated the frameworks for methane storage applications using grand-canonical Monte Carlo simulations, and identified 300 promising methane storage materials. To encourage screening studies of these materials for other applications, all structures and their properties were made available on the Materials Cloud.

Recommended citation: Mercado, Rocío et al. (2018). "In silico design of 2D and 3D covalent organic frameworks for methane storage applications." Chem. Mater. 30(15). 5069-5086. link

Generating carbon schwarzites via zeolite-templating

Published in PNAS, 2018

Here, we develop a theoretical framework for generating zeolite-templated carbons (ZTCs) from any zeolite structure, and use our method to generate a library of ZTCs from all known zeolites. We establish criteria for which zeolites can produce experimentally accessible ZTCs and identify over 10 ZTCs that have not (at the time) been synthesized. We go on to establish a relationship between ZTCs and schwarzites using the triply periodic minimal surface describing the ZTC labyrinths.

Recommended citation: Braun, Efrem et al. (2018). "Generating carbon schwarzites via zeolite-templating." PNAS. 115(35). E8116-E8124. link

Combined nuclear magnetic resonance and molecular dynamics study of methane adsorption in M2(dobdc) metal–organic frameworks

Published in J. Phys. Chem. C., 2019

In this work, we examine the diffusion of methane in MOF-74 through a combination of nuclear magnetic resonance and molecular dynamics simulations. At low gas densities, our results suggest that favorable CH4–CH4 interactions lower the free energy barrier for methane hopping between open metal sites, enhancing the translational motion of methane down the c-axis. Furthermore, we observe that the self-diffusion coefficient of methane is inversely related to the binding energy at the open metal sites, such that diffusion is most rapid in Zn-MOF-74.

Recommended citation: Witherspoon, Velencia J. et al. (2019). "Combined nuclear magnetic resonance and molecular dynamics study of methane adsorption in M2(dobdc) metal–organic frameworks." J. Phys. Chem. C. 123(19). 12286-12295. link

Graph networks for molecular design

Published in ChemRxiv, 2020

(preprint; accepted version above) This work introduces GraphINVENT, a platform developed for graph-based molecular design using graph neural networks (GNNs). GraphINVENT uses a tiered deep neural network architecture to probabilistically generate new molecules a single bond at a time.

Recommended citation: Mercado, Rocío et al. (2020). "Graph networks for molecular design." ChemRxiv. link

Practical notes on building molecular graph generative models

Published in ChemRxiv, 2020

(preprint; accepted version above) This work presents technical notes and tips on developing graph generative models for molecular design. This work stems from the development of GraphINVENT. Technical details that could be of interest to researchers developing their own molecular generative models are discussed, including strategies for designing new models.

Recommended citation: Mercado, Rocío et al. (2020). "Practical notes on building molecular graph generative models." ChemRxiv. link

Molecular representations in AI-driven drug discovery: a review and practical guide

Published in J. Cheminf., 2020

A large number of chemical representations have been developed over the years, their numerosity being due to the fast development of computers and the complexity of producing a representation that encompasses all structural and chemical characteristics. In this mini-review, we present some of the most popular electronic molecular and macromolecular representations used in AI-driven drug discovery.

Recommended citation: David, Laurianne et al. (2020). "Molecular representations in AI-driven drug discovery: a review and practical guide." J. Cheminf. 12(56). link

Comparative study of deep generative models on chemical space coverage

Published in ChemRxiv, 2020

(preprint; accepted version above) This work presents a novel metric for evaluating deep molecular generative models; this new metric compares not only the molecular structures, but also the ring systems and functional groups, reproduced from a reference dataset of 1M structures from GDB-13.

Recommended citation: Zhang, Jie et al. (2020). "Comparative study of deep generative models on chemical space coverage." ChemRxiv. link

Graph networks for molecular design

Published in Mach. Learn.: Sci. Technol., 2020

This work introduces GraphINVENT, a platform developed for graph-based molecular design using graph neural networks (GNNs). GraphINVENT uses a tiered deep neural network architecture to probabilistically generate new molecules a single bond at a time.

Recommended citation: Mercado, Rocío et al. (2020). "Graph networks for molecular design." Mach. Learn.: Sci. Technol. link

Practical notes on building molecular graph generative models

Published in Applied AI Letters, 2020

This work presents technical notes and tips on developing graph generative models for molecular design, as well as an overview of previous work in molecular graph generation. This work stems from the development of GraphINVENT. Technical details that could be of interest to researchers developing their own molecular generative models are discussed, including strategies for designing new models.

Recommended citation: Mercado, Rocío et al. (2020). "Practical notes on building molecular graph generative models." Applied AI Letters. link

Comparative study of deep generative models on chemical space coverage

Published in JCIM, 2021

This work presents a novel and complementary metric for evaluating deep molecular generative models. The performance of seven different molecular generative models was compared by calculating what fraction of the structures, ring systems, and functional groups could be reproduced from the largely unseen reference set when using only a small fraction of GDB-13 for training.

Recommended citation: Zhang, Jie et al. (2021). "Comparative study of deep generative models on chemical space coverage." J. Chem. Inf. Model. link

Exploring graph traversal algorithms in graph-based molecular generation

Published in ChemRxiv, 2021

Here we explore the impact of different graph traversal algorithms on molecular graph generation. We do this by training a graph-based deep molecular generative model to build structures using a node order determined via either a breadth- or depth-first search algorithm.

Recommended citation: Mercado, Rocío et al. (2021) "Exploring graph traversal algorithms in graph-based molecular generation." ChemRxiv. link

Amortized tree generation for bottom-up synthesis planning and synthesizable molecular design

Published in arXiv, 2021

Here, we report an approach which allows us to conduct synthesis planning in a bottom-up manner and design synthesizable molecules by decoding from optimized conditional codes, demonstrating the potential to solve both problems of molecular design and synthesis simultaneously.

Recommended citation: Gao, Wenhao et al. (2021) "Amortized tree generation for bottom-up synthesis planning and synthesizable molecular design." arXiv. link

Exploring graph traversal algorithms in graph-based molecular generation

Published in JCIM, 2021

Here we explore the impact of different graph traversal algorithms on molecular graph generation. We do this by training a graph-based deep molecular generative model to build structures using a node order determined via either a breadth- or depth-first search algorithm.

Recommended citation: Mercado, Rocío et al. (2021) "Exploring graph traversal algorithms in graph-based molecular generation." J. Chem. Inf. Model. link

A transferable Boltzmann generator for small-molecule conformers

Published in ELLIS Machine Learning for Molecule Discovery Workshop, 2021

Here we introduce a transferable generative model for physically realistic conformer ensembles of small molecules. The model is autoregressive and places atoms sequentially using a learned, transferable, and equivariant function implemented using a normalizing flow.

Recommended citation: Viguera Diez, Juan et al. (2021) "A transferable Boltzmann generator for small-molecule conformers." ELLIS ML4Molecules. link

Amortized tree generation for bottom-up synthesis planning and synthesizable molecular design

Published in ICLR 2022, 2022

Molecular design and synthesis planning are two critical steps in the process of molecular discovery that we propose to formulate as a single shared task of conditional synthetic pathway generation. Here, we report an amortized approach to generate synthetic pathways as a Markov decision process conditioned on a target molecular embedding.

Recommended citation: Gao, Wenhao et al. (2022) "Amortized tree generation for bottom-up synthesis planning and synthesizable molecular design." ICLR 2022. link

De novo drug design using reinforcement learning with graph-based deep generative models

Published in JCIM, 2022

Machine learning provides effective computational tools for exploring the chemical space via deep generative models. Here, we propose a new reinforcement learning scheme to fine-tune graph-based deep generative models for de novo molecular design tasks.

Recommended citation: Romeo Atance, Sara et al. (2022) "De novo drug design using reinforcement learning with graph-based deep generative models." J. Chem. Inf. Model. link

De novo PROTAC design using graph-based deep generative models

Published in AI4Science, 2022

PROteolysis TArgeting Chimeras (PROTACs) are an emerging therapeutic modality for degrading a protein of interest (POI) by marking it for degradation by the proteasome. Recent developments in artificial intelligence (AI) suggest that deep generative models can assist with the de novo design of molecules with desired properties, and their application to PROTAC design remains largely unexplored.

Recommended citation: Nori, Divya et al. (2022) "De novo PROTAC design using graph-based deep generative models." NeurIPS 2022 AI4Science Workshop. link

Data sharing in chemistry: lessons learned and a case for mandating structured reaction data

Published in JCIM, 2023

The past decade has seen a number of impressive developments in predictive chemistry and reaction informatics driven by machine learning applications to computer-aided synthesis planning. While many of these developments have been made even with relatively small, bespoke data sets, in order to advance the role of AI in the field at scale, there must be significant improvements in the reporting of reaction data. In this Perspective, we analyze several data curation and sharing initiatives that have seen success in chemistry and molecular biology. We discuss several factors that have contributed to their success and how we can take lessons from these case studies and apply them to reaction data.

Recommended citation: Mercado, Rocio et al. (2023) "Data sharing in chemistry: lessons learned and a case for mandating structured reaction data." J. Chem. Inf. Model. link

Do Chemformers dream of organic matter? Evaluating a transformer model for multi-step retrosynthesis

Published in ChemRxiv, 2023

Here, we trained and evaluated a transformer model, called Chemformer, for retrosynthesis predictions within drug discovery. The proprietary dataset used for training comprised ~18M reactions from literature, patents, and electronic lab notebooks. Chemformer was evaluated for the purpose of both single-step and multi-step retrosynthesis.

Recommended citation: Westerlund, Annie et al. (2023) "Do Chemformers dream of organic matter? Evaluating a transformer model for multi-step retrosynthesis." ChemRxiv. link

Do Chemformers dream of organic matter? Evaluating a transformer model for multi-step retrosynthesis

Published in JCIM, 2024

Here, we trained and evaluated a transformer model, called Chemformer, for retrosynthesis predictions within drug discovery. The proprietary dataset used for training comprised ~18M reactions from literature, patents, and electronic lab notebooks. Chemformer was evaluated for the purpose of both single-step and multi-step retrosynthesis.

Recommended citation: Westerlund, Annie et al. (2024) "Do Chemformers dream of organic matter? Evaluating a transformer model for multi-step retrosynthesis." J. Chem. Inf. Model. link

Modeling PROTAC degradation activity with machine learning

Published in arXiv, 2024

In this work, we present a strategy for curating open-source PROTAC data and an open-source deep learning tool for predicting the degradation activity of novel PROTAC molecules.

Recommended citation: Ribes, Stefano et al. (2024) "Modeling PROTAC degradation activity with machine learning." arXiv. link

Contrastive learning for robust cell annotation and representation from single-cell transcriptomics

Published in bioRxiv, 2024

In this study, we present a novel deep learning approach using contrastive learning and a carefully designed loss function for learning an generalizable embedding space from scRNA-Seq data. We call this model CELLULAR: CELLUlar contrastive Learning for Annotation and Representation.

Recommended citation: Andrekson, Leo et al. (2024) "Contrastive learning for robust cell annotation and representation from single-cell transcriptomics." bioRxiv. link

Modeling PROTAC degradation activity with machine learning

Published in Artificial Intelligence in the Life Sciences, 2024

In this work, we present a strategy for curating open-source PROTAC data and an open-source deep learning tool for predicting the degradation activity of novel PROTAC molecules.

Recommended citation: Ribes, Stefano et al. (2024) "Modeling PROTAC degradation activity with machine learning." Artificial Intelligence in the Life Sciences. link

recent_talks

upcoming_talks