What question did this study set out to answer?

To develop a computational framework for predicting protein carbonylation sites across lysine, arginine, threonine, and proline residues by analyzing substrate sequence and physicochemical properties.

March 1, 2017BMC BioinformaticsOpen Access

Investigation and identification of protein carbonylation sites based on position-specific amino acid composition and physicochemical features

View Full Paper

Ask AI

Bookmark

Authors

SWShun-Long WengNational Yang Ming Chiao Tung University KHKai‐Yao HuangMackay Memorial Hospital FKFergie Joanda Kaunang

Discussion

Member takes

Implication

Machine learning study demonstrates accurate prediction of human protein carbonylation sites using hybrid sequence features, highlighting positional amino acid enrichment during oxidative stress.

Key Points

To develop a computational framework for predicting protein carbonylation sites across lysine, arginine, threonine, and proline residues by analyzing substrate sequence and physicochemical properties.
Extracted 226 non-redundant human carbonylated proteins containing 307 lysine (K), 126 arginine (R), 128 threonine (T), and 129 proline (P) carbonylation sites from published literature.
Trained and evaluated models using three machine learning algorithms across multiple sequence-based feature encodings (AAC, AAPC, PSSM, PWM, ASA, and AAindex) with five-fold cross-validation and an independent test set.
Positively charged residues (lysine and arginine) were significantly enriched in regions surrounding carbonylation sites, distinguishing modified from non-modified residues.
A hybrid predictive model combining positional weighted matrix, amino acid composition, and AAindex achieved Matthews correlation coefficients of 0.432 for K, 0.472 for R, 0.443 for T, and 0.467 for P residues.
The hybrid feature model yielded higher predictive accuracy than an existing carbonylation prediction tool when evaluated on an independent test dataset.

Cite This Study

Weng et al. (2017) studied this question.

synapsesocial.com/papers/6a1bf4bcd54006be995f4ba4 https://doi.org/10.1186/s12859-017-1472-8

View Full Paper

Ask AI

Bookmark

March 1, 2017BMC BioinformaticsOpen Access

Investigation and identification of protein carbonylation sites based on position-specific amino acid composition and physicochemical features

View Full Paper

Ask AI

Bookmark

Authors

SWShun-Long WengNational Yang Ming Chiao Tung University KHKai‐Yao HuangMackay Memorial Hospital FKFergie Joanda Kaunang

Discussion

Member takes

Implication

Machine learning study demonstrates accurate prediction of human protein carbonylation sites using hybrid sequence features, highlighting positional amino acid enrichment during oxidative stress.

Key Points

To develop a computational framework for predicting protein carbonylation sites across lysine, arginine, threonine, and proline residues by analyzing substrate sequence and physicochemical properties.
Extracted 226 non-redundant human carbonylated proteins containing 307 lysine (K), 126 arginine (R), 128 threonine (T), and 129 proline (P) carbonylation sites from published literature.
Trained and evaluated models using three machine learning algorithms across multiple sequence-based feature encodings (AAC, AAPC, PSSM, PWM, ASA, and AAindex) with five-fold cross-validation and an independent test set.
Positively charged residues (lysine and arginine) were significantly enriched in regions surrounding carbonylation sites, distinguishing modified from non-modified residues.
A hybrid predictive model combining positional weighted matrix, amino acid composition, and AAindex achieved Matthews correlation coefficients of 0.432 for K, 0.472 for R, 0.443 for T, and 0.467 for P residues.
The hybrid feature model yielded higher predictive accuracy than an existing carbonylation prediction tool when evaluated on an independent test dataset.

Cite This Study

Weng et al. (2017) studied this question.

synapsesocial.com/papers/6a1bf4bcd54006be995f4ba4 https://doi.org/10.1186/s12859-017-1472-8

View Full Paper

Ask AI

Bookmark