Characterization and identification of lysine crotonylation sites based on machine learning method on both plant and mammalian

Abstract Lysine crotonylation (Kcr) is a type of protein post-translational modification (PTM), which plays important roles in a variety of cellular regulation and processes. Several methods have been proposed for the identification of crotonylation. However, most of these methods can predict effici...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Rulan Wang, Zhuo Wang, Hongfei Wang, Yuxuan Pang, Tzong-Yi Lee
Formato: article
Lenguaje:EN
Publicado: Nature Portfolio 2020
Materias:
R
Q
Acceso en línea:https://doaj.org/article/a853eda2aeaf4504982c8c920cfc2d05
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:a853eda2aeaf4504982c8c920cfc2d05
record_format dspace
spelling oai:doaj.org-article:a853eda2aeaf4504982c8c920cfc2d052021-12-02T12:33:46ZCharacterization and identification of lysine crotonylation sites based on machine learning method on both plant and mammalian10.1038/s41598-020-77173-02045-2322https://doaj.org/article/a853eda2aeaf4504982c8c920cfc2d052020-11-01T00:00:00Zhttps://doi.org/10.1038/s41598-020-77173-0https://doaj.org/toc/2045-2322Abstract Lysine crotonylation (Kcr) is a type of protein post-translational modification (PTM), which plays important roles in a variety of cellular regulation and processes. Several methods have been proposed for the identification of crotonylation. However, most of these methods can predict efficiently only on histone or non-histone protein. Therefore, this work aims to give a more balanced performance in different species, here plant (non-histone) and mammalian (histone) are involved. SVM (support vector machine) and RF (random forest) were employed in this study. According to the results of cross-validations, the RF classifier based on EGAAC attribute achieved the best predictive performance which performs competitively good as existed methods, meanwhile more robust when dealing with imbalanced datasets. Moreover, an independent test was carried out, which compared the performance of this study and existed methods based on the same features or the same classifier. The classifiers of SVM and RF could achieve best performances with 92% sensitivity, 88% specificity, 90% accuracy, and an MCC of 0.80 in the mammalian dataset, and 77% sensitivity, 83% specificity, 70% accuracy and 0.54 MCC in a relatively small dataset of mammalian and a large-scaled plant dataset respectively. Moreover, a cross-species independent testing was also carried out in this study, which has proved the species diversity in plant and mammalian.Rulan WangZhuo WangHongfei WangYuxuan PangTzong-Yi LeeNature PortfolioarticleMedicineRScienceQENScientific Reports, Vol 10, Iss 1, Pp 1-12 (2020)
institution DOAJ
collection DOAJ
language EN
topic Medicine
R
Science
Q
spellingShingle Medicine
R
Science
Q
Rulan Wang
Zhuo Wang
Hongfei Wang
Yuxuan Pang
Tzong-Yi Lee
Characterization and identification of lysine crotonylation sites based on machine learning method on both plant and mammalian
description Abstract Lysine crotonylation (Kcr) is a type of protein post-translational modification (PTM), which plays important roles in a variety of cellular regulation and processes. Several methods have been proposed for the identification of crotonylation. However, most of these methods can predict efficiently only on histone or non-histone protein. Therefore, this work aims to give a more balanced performance in different species, here plant (non-histone) and mammalian (histone) are involved. SVM (support vector machine) and RF (random forest) were employed in this study. According to the results of cross-validations, the RF classifier based on EGAAC attribute achieved the best predictive performance which performs competitively good as existed methods, meanwhile more robust when dealing with imbalanced datasets. Moreover, an independent test was carried out, which compared the performance of this study and existed methods based on the same features or the same classifier. The classifiers of SVM and RF could achieve best performances with 92% sensitivity, 88% specificity, 90% accuracy, and an MCC of 0.80 in the mammalian dataset, and 77% sensitivity, 83% specificity, 70% accuracy and 0.54 MCC in a relatively small dataset of mammalian and a large-scaled plant dataset respectively. Moreover, a cross-species independent testing was also carried out in this study, which has proved the species diversity in plant and mammalian.
format article
author Rulan Wang
Zhuo Wang
Hongfei Wang
Yuxuan Pang
Tzong-Yi Lee
author_facet Rulan Wang
Zhuo Wang
Hongfei Wang
Yuxuan Pang
Tzong-Yi Lee
author_sort Rulan Wang
title Characterization and identification of lysine crotonylation sites based on machine learning method on both plant and mammalian
title_short Characterization and identification of lysine crotonylation sites based on machine learning method on both plant and mammalian
title_full Characterization and identification of lysine crotonylation sites based on machine learning method on both plant and mammalian
title_fullStr Characterization and identification of lysine crotonylation sites based on machine learning method on both plant and mammalian
title_full_unstemmed Characterization and identification of lysine crotonylation sites based on machine learning method on both plant and mammalian
title_sort characterization and identification of lysine crotonylation sites based on machine learning method on both plant and mammalian
publisher Nature Portfolio
publishDate 2020
url https://doaj.org/article/a853eda2aeaf4504982c8c920cfc2d05
work_keys_str_mv AT rulanwang characterizationandidentificationoflysinecrotonylationsitesbasedonmachinelearningmethodonbothplantandmammalian
AT zhuowang characterizationandidentificationoflysinecrotonylationsitesbasedonmachinelearningmethodonbothplantandmammalian
AT hongfeiwang characterizationandidentificationoflysinecrotonylationsitesbasedonmachinelearningmethodonbothplantandmammalian
AT yuxuanpang characterizationandidentificationoflysinecrotonylationsitesbasedonmachinelearningmethodonbothplantandmammalian
AT tzongyilee characterizationandidentificationoflysinecrotonylationsitesbasedonmachinelearningmethodonbothplantandmammalian
_version_ 1718393877686124544