Computational Tools for Quantifying Poaceae Pollen Diversity and Photosynthetic Pathway Composition (C3/C4) from Superresolution Images

Fuente: Zenodo
Guardado en:
Detalles Bibliográficos
Autores principales: Adaimé, Marc-Élie, Kong, Shu, Punyasena, Surangi W.
Formato: Recurso digital
Publicado: Zenodo 2024
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866902284494962688
author Adaimé, Marc-Élie
Kong, Shu
Punyasena, Surangi W.
author_facet Adaimé, Marc-Élie
Kong, Shu
Punyasena, Surangi W.
contents <p>This release includes the codebase and models supporting the study titled "<em>Reconstructing the diversity dynamics of paleo-grasslands using deep learning on superresolution images of fossil Poaceae pollen.</em>" </p> <p>The pipeline enables: </p> <ul> <li>Classification of superresolution images of grass (Poaceae) pollen grains using convolutional neural networks (CNNs) trained on 60 extant species. </li> <li>Extraction of high-level morphological features that serve as proxies for taxonomic diversity, quantified <em>via</em> Shannon entropy over the probability distribution of morphological variability within each community.</li> <li>Discrimination between C3 and C4 grass species based solely on pollen morphology, using a random forest classifier. </li> <li>Temporal reconstruction of past diversity and C3 <em>vs</em>. C4 composition across different time periods, and their comparison to independent environmental records, including atmospheric CO2, temperature, precipitation, and fire history. </li> </ul> <p>The use of Shannon entropy as a diversity proxy is validated through ecological simulations, which are also included in this repository. </p>
format Recurso digital
id zenodo_https___doi_org_10_5281_zenodo_13756262
institution Zenodo
language
publishDate 2024
publisher Zenodo
record_format zenodo
spellingShingle Computational Tools for Quantifying Poaceae Pollen Diversity and Photosynthetic Pathway Composition (C3/C4) from Superresolution Images
Adaimé, Marc-Élie
Kong, Shu
Punyasena, Surangi W.
Poaceae
Pollen Morphology
Deep Learning
Machine Learning
Random Forest
Convolutional Neural Networks
Photosynthetic Pathway
Superresolution Microscopy
Shannon Entropy
Grassland
Grass Evolution
Biological Community Simulations
Computational Biology
<p>This release includes the codebase and models supporting the study titled "<em>Reconstructing the diversity dynamics of paleo-grasslands using deep learning on superresolution images of fossil Poaceae pollen.</em>" </p> <p>The pipeline enables: </p> <ul> <li>Classification of superresolution images of grass (Poaceae) pollen grains using convolutional neural networks (CNNs) trained on 60 extant species. </li> <li>Extraction of high-level morphological features that serve as proxies for taxonomic diversity, quantified <em>via</em> Shannon entropy over the probability distribution of morphological variability within each community.</li> <li>Discrimination between C3 and C4 grass species based solely on pollen morphology, using a random forest classifier. </li> <li>Temporal reconstruction of past diversity and C3 <em>vs</em>. C4 composition across different time periods, and their comparison to independent environmental records, including atmospheric CO2, temperature, precipitation, and fire history. </li> </ul> <p>The use of Shannon entropy as a diversity proxy is validated through ecological simulations, which are also included in this repository. </p>
title Computational Tools for Quantifying Poaceae Pollen Diversity and Photosynthetic Pathway Composition (C3/C4) from Superresolution Images
topic Poaceae
Pollen Morphology
Deep Learning
Machine Learning
Random Forest
Convolutional Neural Networks
Photosynthetic Pathway
Superresolution Microscopy
Shannon Entropy
Grassland
Grass Evolution
Biological Community Simulations
Computational Biology
url https://doi.org/10.5281/zenodo.13756262