<i>ChemML</i>: A machine learning and informatics program package for the analysis, mining, and modeling of chemical and materials data

Haghighatlari, Mojtaba; Vishwakarma, Gaurav; Altarawy, Doaa; Subramanian, R.; Kota, Bhargava Urala; Sonpal, Aditya; Setlur, Srirangaraj; Hachmann, Johannes

doi:10.1002/wcms.1458

Cited by 48 publications

(30 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In terms of feature representation, systematic development of new descriptors, standardization of their evaluation, and easier accessibility via user interfaces (e.g., Python libraries) are necessary to establish their long-term development. 60 A transparent and sustained study of feature representation would involve researchers with variety of domain knowledge and expertise to accelerate future developments.…”

Section: Llmentioning

confidence: 99%

“…16 The Bigger Picture Scarce and sparse chemical datasets. The number of unique small molecules is practically infinite, with number estimates of possible synthesizable small molecules ranging from 10 24 -10 60 . Machine learning could expand coverage of chemical space-for instance, in the design, synthesis, and development stages of drugsthat traditionally are resourceintensive tasks.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Learning to Make Chemical Predictions: The Interplay of Feature Representation, Data, and Machine Learning Methods

Haghighatlari

Heidar‐Zadeh

et al. 2020

Chem

Self Cite

View full text Add to dashboard Cite

Recently, supervised machine learning has been ascending in providing new predictive approaches for chemical, biological, and materials sciences applications. In this Perspective, we focus on the interplay of machine learning methods with the chemically motivated descriptors and the size and type of datasets needed for molecular property prediction. Using nuclear magnetic resonance chemical shift prediction as an example, we demonstrate that success is predicated on the choice of feature extracted or real-space representations of chemical structures, whether the molecular property data are abundant and/or experimentally or computationally derived, and how these together will influence the correct choice of popular machine learning methods drawn from deep learning, random forests, or kernel methods.

show abstract

Section: Llmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Learning to Make Chemical Predictions: The Interplay of Feature Representation, Data, and Machine Learning Methods

Haghighatlari

Heidar‐Zadeh

et al. 2020

Chem

Self Cite

View full text Add to dashboard Cite

show abstract

“…Initial molecular geometries for all 962 P-molecules were obtained from SMILES codes through a Python script utilizing the RDKit (RDK, 2000), ChemML (Haghighatlari et al, 2020), and ChemCoord (Weser, 2017) libraries.…”

Section: Methodsmentioning

confidence: 99%

Computational Infrared Spectroscopy of 958 Phosphorus-Bearing Molecules

Trujilo

Syme

Rowell

et al. 2021

Front. Astron. Space Sci.

View full text Add to dashboard Cite

Phosphine is now well-established as a biosignature, which has risen to prominence with its recent tentative detection on Venus. To follow up this discovery and related future exoplanet biosignature detections, it is important to spectroscopically detect the presence of phosphorus-bearing atmospheric molecules that could be involved in the chemical networks producing, destroying or reacting with phosphine. We start by enumerating phosphorus-bearing molecules (P-molecules) that could potentially be detected spectroscopically in planetary atmospheres and collecting all available spectral data. Gaseous P-molecules are rare, with speciation information scarce. Very few molecules have high accuracy spectral data from experiment or theory; instead, the best current spectral data was obtained using a high-throughput computational algorithm, RASCALL, relying on functional group theory to efficiently produce approximate spectral data for arbitrary molecules based on their component functional groups. Here, we present a high-throughput approach utilizing established computational quantum chemistry methods (CQC) to produce a database of approximate infrared spectra for 958 P-molecules. These data are of interest for astronomy and astrochemistry (importantly identifying potential ambiguities in molecular assignments), improving RASCALL's underlying data, big data spectral analysis and future machine learning applications. However, this data will probably not be sufficiently accurate for secure experimental detections of specific molecules within complex gaseous mixtures in laboratory or astronomy settings. We chose the strongly performing harmonic ωB97X-D/def2-SVPD model chemistry for all molecules and test the more sophisticated and time-consuming GVPT2 anharmonic model chemistry for 250 smaller molecules. Limitations to our automated approach, particularly for the less robust GVPT2 method, are considered along with pathways to future improvements. Our CQC calculations significantly improve on existing RASCALL data by providing quantitative intensities, new data in the fingerprint region (crucial for molecular identification) and higher frequency regions (overtones, combination bands), and improved data for fundamental transitions based on the specific chemical environment. As the spectroscopy of most P-molecules have never been studied outside RASCALL and this approach, the new data in this paper is the most accurate spectral data available for most P-molecules and represent a significant advance in the understanding of the spectroscopic behavior of these molecules.

show abstract

“…Los 28 textos que abordan la CD y la CI y recibieron citas, han sido elaborados de manera colectiva (67.2 %) o individual (32.8 %). Por un lado, aquellos artículos con más integrantes participantes y citas recibidas son: "Geospatial Data Management Research: Progress and Future Directions" (Breunig et al, 2020) con nueve participantes y 1 cita recibida; "ChemML: A machine learning La informetría y el análisis del discurso aplicados a la producción científica en la Ciencia de Datos y Ciencia de la información Celso Martínez Musiño 9 and informatics program package for the analysis, mining, and modeling of chemical and materials data" (Haghighatlari et al, 2020), con ocho investigadores; 5 citas recibidas, y; "Reproducible research and GIScience: an evaluation using AGILE conference papers" (Nüst et al, 2018) Morriello (2020), redactados en portugués, francés, japonés e italiano. Aunque hay un predominio de la lengua inglesa como medio natural de la comunicación científica, se observa la presencia de idiomas distintos que cumplen con las exigencias y requerimientos normativos de las revistas y las bases de datos especializadas.…”

Section: Autoría E Idiomaunclassified

La informetría y el análisis del discurso aplicados a la producción científica en la Ciencia de Datos y Ciencia de la información

Musiño¹

2021

e-Ciencias de la Información

View full text Add to dashboard Cite

Objetivo. Investigar la producción científica en la intersección temática de la Ciencia de Datos (CD) y la Ciencia de la Información (CI). Método. Estudio informétrico, descriptivo, y de primera incursión, en el análisis del discurso escrito de los textos académicos incluidos en la Web of Science (WoS), periodo 1900 al 6 de noviembre de 2020, y cuya cobertura de búsqueda fue en las bases de datos: SCI-Expanded, SSCI, A&HCI, ESCI, CPCI-S, CPCI-SSH, BKCI-S y BKCI-SSH. Resultados. Se recuperaron y analizaron 49 documentos representados en 38 artículos, 7 textos de memorias de congresos, 2 capítulos de libro y 2 materiales editoriales. El conjunto de las investigaciones que tratan el tema de la CD y la CI sumaron 128 citas, 2.6 citas por documento e índice H: 7. Discusión. Conceptualmente, se encontró que para la CD y la CI su origen son los “datos” y que ambas disciplinas son predominantemente de carácter práctico. En aquellas investigaciones con mayor visibilidad hay participación multiautoral. La CD y la CI son áreas del conocimiento recientes en las cuales las tecnologías de la información son indispensables para el análisis de grandes cantidades de datos e información. Conclusiones. La CD y CI tienen un carácter intra y multi y transdisciplinar y se caracterizan por utilizar las tecnologías de la información para el análisis de grandes cantidades de datos e información.

show abstract

ChemML: A machine learning and informatics program package for the analysis, mining, and modeling of chemical and materials data

Cited by 48 publications

References 40 publications

Learning to Make Chemical Predictions: The Interplay of Feature Representation, Data, and Machine Learning Methods

Learning to Make Chemical Predictions: The Interplay of Feature Representation, Data, and Machine Learning Methods

Computational Infrared Spectroscopy of 958 Phosphorus-Bearing Molecules

La informetría y el análisis del discurso aplicados a la producción científica en la Ciencia de Datos y Ciencia de la información

Contact Info

Product

Resources

About