Amino acid sequence

Balance data

Group	Type	Literature	Year	Data	Acc		URL
Function Unequal length sequence
	DNA-Binding Proteins	TargetDBP plus: Enhancing the Performance of Identifying DNABinding Proteins via Weighted Convolutional Features	2021	Download ↓ *UniSwiss-Tr: DBPs: 4500; Non-DBPs: 4500; UniSwiss-Tst:* DBPs: 381; Non-DBPs: 381;	jackknife test 86.84%	Independent testing 85.83%	TargetDBP plus
	DNA-Binding Proteins	Use Chou’s 5-Step Rule to Predict DNA-Binding Proteins with Evolutionary Information	2020	Download ↓ Training dataset-PDB1075: DBPs: 525; Non-DBPs: 525; Independent testing-PDB186: DBPs: 93; Non-DBPs: 93;	jackknife test PDB1075: 86.05%	Independent testing PDB186: 75.30%	PSSMEI
	Gram-negative bacteria	BastionHub: a universal platform for integrating and analyzing substrates secreted by Gram-negative bacteria	2021	Download ↓ T1SS: 195; T2SS: 83; T3SS: 1194; T4SS: 713; T6SS: 181;	/		BastionHub
	Bacterial Type III Secreted Effectors	iT3SE-PX: Identification of Bacterial Type III Secreted Effectors Using PSSM Profiles and XGBoost Feature Selection	2021	Download ↓ Training dataset: P-379/ N-1112; Testing dataset: P-108/ N-108;	5-CV 0.967±0.002	Independent testing 0.963±0.034	iT3SE-PX
	Bacterial Type III Secreted Effectors	T3SEpp: an Integrated Prediction Pipeline for Bacterial Type III Secreted Effectors	2020	Download ↓ Training dataset: P-309/ N-310; Testing dataset: P-42/ N-34;	10-CV 0.941±0.011	Independent testing 0.83	T3SEpp
	Soluble protein	NetSolP: predicting protein solubility in Escherichia coli using language models	2022	Download ↓ Training dataset: P-5718/ N-5718; Testing dataset: P-1550/ N-1550;	5-CV 0.70±0.02	Independent testing 0.728	NetSolP
	Soluble protein	SoluProt: prediction of soluble protein expression in Escherichia coli	2021		5-CV 76.0%	Independent testing 58.5%	SoluProt
	Secretory Proteins	CRCF: A Method of Identifying Secretory Proteins of Malaria Parasites	2021	Download ↓ Secretory proteins: 252 Non-secretory protein: 252	10-CV 97.8%		CRCF
	Cancerlectin	The accurate prediction and characterization of cancerlectin by a combined machine learning and GO analysis	2021	Download ↓ Cancerlectin sequences: 176; Non-Cancerlectin sequences: 227; Lectins sequences: 462; Non-Lectins sequences: 435;	5-CV Cancerlectin: 98.54% Lectin: 95.38%		/
	Cancerlectin	Identification of Cancerlectins Using Support Vector Machines With Fusion of G-Gap Dipeptide	2020	Download ↓ Cancerlectin sequences: 178 Non-Cancerlectin sequences: 226	10-CV 83.91%		Cancerlectins
	Phage virion protein	Review and comparative analysis of machine learning-based phage virion protein identification methods	2020	Download ↓ Training dataset: P-99/ N-208 Testing dataset: P-30/ N-60	jackknife test 0.8958	Independent testing 0.7980	SVM-PVPData
	Cell wall lytic enzymes	CWLy-RF: A novel approach for identifying cell wall lyases based on random forest classifier	2021	Download ↓ Lyases: 68 Non-Lyases: 307	10-CV 96.09%		/
	Cell wall lytic enzymes	CWLy-SVM: A support vector machine-based tool for identifying cell wall lytic enzymes	2020	Download ↓ Lyases: 68 Non-Lyases: 307	jackknife test 0.955		CWLy-SVM
	Thermophilic Proteins	A Method for Prediction of Thermophilic Protein Based on Reduced Amino Acids and Mixed Features	2020	Download ↓ Thermophilic proteins: 915 Non-thermophilic proteins: 793	10-CV 98.02%		Tool
		Discrimination of Thermophilic Proteins and Non-thermophilic Proteins Using Feature Dimension Reduction	2020		cross-validation 96.02%		/
		Prediction of Thermophilic Proteins Using Voting Algorithm	2019		10-CV 93.03%		Tool
	Major Histocompatibility Complex	ELM-MHC: An Improved MHC Identification Method with Extreme Learning Machine Algorithm	2019	Download ↓ Dataset1: Training dataset: MHC protein: 4370; Non-MHC protein: 4370; Testing dataset: MHC protein: 2342; Non-MHC protein: 2342; Dataset2: MHC I: 3350; MHC II: 3362;	10-CV Dataset1: 91.66%; Dataset2: 94.224%	Independent testing Dataset1: 93.17%	ELM-MHC
	Amyloid Proteins	iAMY-SCM: Improved prediction and analysis of amyloid proteins using a scoring card method with propensity scores of dipeptides	2021	Download ↓ Amyloid proteins: 165 Non_amyloid proteins: 382 (ps: 80% Training dataset, 20% Testing dataset)	10-CV 0.895	Independent testing 0.827	iAMY-SCM
		PredAmyl-MLP: Prediction of Amyloid Proteins Using Multilayer Perceptron	2020		10-CV 91.59%		PredAmyl-MLP
		RFAmyloid: A Web Server for Predicting Amyloid Proteins	2018		10-CV 89.19%	Independent testing 89.19%	RFAmyloid
Modification Equal length sequence	Multiple PTMs	Prediction and analysis of multiple protein lysine modified sites based on conditional wasserstein generative adversarial networks	2021	Download ↓ S1(Ace): 3114; S2(Glyca): 1399; S3(Malon): 1224; S4(Meth): 1147; S5(Succ): 1645; S6(Sumo): 1174; S7(Ubiq): 3185; (ps: 4/5 of all samples were used as training samples and the other 1/5 was utilized as an independent test set)	10-CV 0.8589	Independent testing 0.8549	MultiLyGAN
	Acetylation	DNNAce: Prediction of prokaryote lysine acetylation sites through deep neural networks with multi-information fusion	2020	Download ↓ Archaea: Training dataset: P-193/ N-193; Testing dataset: P-21/ N-21; B.subtilis: Training dataset: P-1040/ N-1040; Testing dataset: P-115/ N-115; C.glutamicum: Training dataset: P-1021/ N-1021; Testing dataset: P-113/ N-113; E.amylovora: Training dataset: P-95/ N-95; Testing dataset: P-10/ N-10; E.coli: Training dataset: P-1919/ N-1919; Testing dataset: P-213/ N-213; G.kaustophilus: Training dataset: P-189/ N-189; Testing dataset: P-21/ N-21; M.tuberculosis: Training dataset: P-866/ N-866; Testing dataset: P-96/ N-96; S.typhimuricum: Training dataset: P-174/ N-174; Testing dataset: P-19/ N-19; V.parahemolvticus: Training dataset: P-1065/ N-1065; Testing dataset: P-118/ N-118; ProAcePred Negative data using this dataset	10-CV Archaea: 84.74%; B.subtilis: 73.89%; C.glutamicum: 75.38%; E.amylovora: 96.89%; E.coli: 63.08%; G.kaustophilus: 89.15%; M.tuberculosis: 76.62%; S.typhimuricum: 90.51%; V.parahemolvticus: 75.46%;	Independent testing Archaea: 90.00%; B.subtilis: 98.26%; C.glutamicum: 92.88%; E.amylovora: 90.00%; E.coli: 86.18%; G.kaustophilus: 97.50%; M.tuberculosis: 96.44%; S.typhimuricum: 95.00%; V.parahemolvticus: 94.02%;	DNNAce
		ProAcePred: prokaryote lysine acetylation sites prediction based on elastic net feature optimization	2018		10-CV Archaea: 0.9; B.subtilis: 0.796; C.glutamicum: 0.8; E.amylovora: 0.983; E.coli: 0.69; G.kaustophilus: 0.897; M.tuberculosis: 0.834; S.typhimuricum: 0.874; V.parahemolvticus: 0.802;	Independent testing Archaea: 0.81; B.subtilis: 0.952; C.glutamicum: 0.872; E.amylovora: 0.9; E.coli: 0.899; G.kaustophilus: 0.881; M.tuberculosis: 0.88; S.typhimuricum: 0.816; V.parahemolvticus: 0.869;	ProAcePred^∗
		A deep learning method to more accurately recall known lysine acetylation sites	2019	Large Data Download ↓ Training dataset: P-12886/N-12886 Testing dataset: P-3221/ N-3221	10-CV 84.95%	Independent testing 84.87%	DeepAcet
	Hydroxylation	iHyd-LysSite (EPSV): Identifying Hydroxylysine Sites in Protein Using Statistical Formulation by Extracting Enhanced Position and Sequence Variant Feature Technique	2020	Download ↓ Hydroxylysine sites: 185 Non-Hydroxylysine sites: 497	jackknife test 97.24%		/
	Malonylation	Mal-Prec: computational prediction of protein Malonylation sites via machine learning based feature integration : Malonylation site prediction	2020	Download ↓ Malonylation sites: 1735 Non-malonylation sites: 1735 (ps: 80% Training dataset, 20% Testing dataset)	5-CV 91.24%	Independent testing 90.65%	Mal-Prec
	Malonylation	Computational analysis and prediction of lysine malonylation sites by exploiting informative features in an integrative machine-learning framework	2019	Download ↓ H. sapiens Training dataset: P-3585/ N-3585; Testing dataset: P-300/ N-300; M. musculus Training dataset: P-2606/ N-2606; Testing dataset: P-300/ N-300; E. coli Training dataset: P-1453/ N-1453; Testing dataset: P-100/ N-100;	10-CV E. coli: 0.801±0.002 M. musculus: 0.825±0.002 H. sapiens: 0.835±0.002	Independent testing E. coli: 0.845 M. musculus: 0.833 H. sapiens: 0.860	KMAL-SP
	Methylation	Computational Prediction of Protein Arginine Methylation Based on Composition−Transition−Distribution Features	2020	Download ↓ Single-methylarginine: 1465 double-methylarginine: 474 Negative samples: 39980	10-CV Single: 82.1% Double: 82.5%		Code
	Methylation	Fast prediction of protein methylation sites using a sequence-based feature selection technique	2017	Download ↓ Methylation site R Dataset-1: P-185/ N-185; Dataset-2: P-185/ N-185; Dataset-3: P-185/ N-185; Dataset-4: P-185/ N-185; Dataset-5: P-185/ N-185; Dataset-6: P-185/ N-186; Dataset-7: P-185/ N-185; Methylation site K Dataset-1: P-226/ N-217; Dataset-2: P-226/ N-217; Dataset-3: P-226/ N-217; Dataset-4: P-226/ N-217; Dataset-5: P-226/ N-218; Dataset-6: P-226/ N-217; Dataset-7: P-226/ N-217;	jackknife test Methylation site R Dataset-1: 80.3%; Dataset-2: 80.3%; Dataset-3: 78.4%; Dataset-4: 80.8%; Dataset-5: 82.2%; Dataset-6: 82.7%; Dataset-7: 80.5%; Average: 80.7%; Methylation site K Dataset-1: 67.7%; Dataset-2: 68.4%; Dataset-3: 72.7%; Dataset-4: 70.0%; Dataset-5: 69.0%; Dataset-6: 69.3%; Dataset-7: 72.5%; Average: 69.9%;		MePred-RF
	Palmitoylation	SPalmitoylC-PseAAC: A sequence-based model developed via Chou's 5-steps rule and general PseAAC for identifying S-palmitoylation sites in proteins	2019	Download ↓ S-palmitoylation sites: 436 Non-S-palmitoylation sites: 500	Self-consistency testing 99.79%	10-CV 97.22%	SPalmitoylC-PseAAC^∗
	Phosphorylation	DeepPPSite: A deep learning-based model for analysis and prediction of phosphorylation sites using efficient sequence information	2021	Download ↓ Training dataset: S: P-4316/ N-4316; T: P-1551/ N-1551; Y: P-553/ N-553; Testing dataset: S: P-2773/ N-17118; T: P-941/ N-6258; Y: P-210/ N-1296;	10-CV S: 80.38%; T: 80.01%; Y: 77.76%;	Independent testing S: 78.91%; T: 84.81%; Y: 82.73%;	DeepPPSite
		DeepIPs: comprehensive assessment and computational identification of phosphorylation sites of SARS-CoV-2 infection using a deep learning-based approach	2021	Download ↓ Training dataset: S/T: P-4308/ N-4308; Y: P-81/ N-81; Testing dataset: S/T: P-1079/ N-1079; Y: P-21/ N-21	5-CV S/T: 80.45% Y: 75.22%	Independent testing S/T: 80.63% Y: 83.33%	DeepIPs
		iPhoPred: A Predictor for Identifying Phosphorylation Sites in Human Protein	2019	Download ↓ SerD: P-300/ N-300; TyrD: P-110/ N-110; ThrD: P-100/ N-100;	jackknife test SerD: AUC 0.904 TyrD: AUC 0.992 ThrD: AUC 0.990		iPhoPred
		PhosPred-RF: a novel sequence-based predictor for phosphorylation sites using sequential information only	2017	Download ↓ Training dataset: S-type: P-4316/ N-4316; T-type: P-1551/ N-1551; Y-type: P-553/ N-553; Testing dataset: S-type: P-2273/ N-17618; T-type: P-941/ N-6258; Y-type: P-296/ N-1210;	10-CV S-type: AUC 0.851 T-type: AUC 0.818 Y-type: AUC 0.761	Independent testing S-type: AUC 0.715 T-type: AUC 0.683 Y-type: AUC 0.654	PhosPred-RF
	Pupylation	PUP-Fuse: Prediction of Protein Pupylation Sites by Integrating Multiple Sequence Representations	2021	Download ↓ Training dataset: Pupylated lysine: 186; Non-pupylated lysine: 186; Testing dataset: Pupylated lysine: 87; Non-pupylated lysine: 191;	10-CV 0.884	Independent testing 0.82	PUP-Fuse
	Succinylation	LSTMCNNsucc: A Bidirectional LSTM and CNN-Based Deep Learning Method for Predicting Lysine Succinylation Sites	2021	Download ↓ Training dataset: Succinylation sites: 6512; Non-Succinylation sites: 6512; Testing dataset: Succinylation sites: 1479; Non-Succinylation sites: 16457;	10-CV 79.9%	NA	LSTMCNNsucc
	Tyrosine sulfation	iSulfoTyr-PseAAC: Identify Tyrosine Sulfation Sites by Incorporating Statistical Moments via Chou’s 5-steps Rule and Pseudo Components	2019	Download ↓ Training dataset: P-200/ N-420 Testing dataset: P-80/ N-80	10-CV 94.2%	Independent testing 85.63%	/
	Ubiquitination	DeepTL-Ubi: a novel deep transfer learning method for effectively predicting ubiquitination sites of multiple species	2021	Download ↓ H.sapiens: P-31162/ N-31162; M.musculus: P-7746/ N-7748; S.cerevisiae: P-3506/ N-3506; R.norvegicus: P-1226/ N-1226; A.nidulans: P-2299/ N-2299; A.thaliana: P-586/ N-587; T.gondii: P-424/ N-424; O.sativa: P-308/ N-308; (ps: Training dataset 90% and Independent test dataset 10%)	H. sapiens: 0.578; M.musculus: 0.604; R.norvegicus: 0.578; S.cerevisiae: 0.593; A. thaliana: 0.525; O. sativa: 0.500; T.gondii: 0.538; A. nidulans: 0.636;	Independent testing H.sapiens: AUC 0.753; M.musculus: AUC 0.789; R.norvegicus: AUC 0.72; S.cerevisiae: AUC 0.772; T.gondii: AUC 0.824; A.thaliana: AUC 0.814;	DeepTL-Ubi
	Ubiquitination	UbiSitePred: A novel method for improving the accuracy of ubiquitination sites prediction by using LASSO to select the optimal Chou's pseudo components	2019	Download ↓ Set1: P-150/ N-150; Set2: P-3418/ N-3418; Set3: P-6117/ N-6117;	5-CV Set1: 98.33%; Set2: 81.12%; Set3: 76.90%		UbiSitePred

Note: 5-CV：5-fold cross validation; 10-CV: 10-fold cross validation; P: Positive samples; N: Negative samples;
^∗Web-server/Code is not available at the time of writing

Imbalance data

Group	Type	Literature	Year	Data	Imbalance Algorithms	Acc		URL
Function Unequal length sequence	Antioxidant proteins	ORS-Pred: An optimized reduced scheme-based identifier for antioxidant proteins	2021	Download ↓ Antioxidant proteins: 253 Non-antioxidant proteins:1552	SMOTE	10-CV 97.00%		ORS-Pred
		Prediction of antioxidant proteins using hybrid feature representation method and random forest	2020	Download ↓ Antioxidant proteins: 434 Non-antioxidant proteins: 1550	/	8-CV 0.982		/
		AOPs-SVM: A Sequence-Based Classifier of Antioxidant Proteins Using a Support Vector Machine	2019	Download ↓ Antioxidant proteins: 253 Non-antioxidant proteins: 1552	/	jackknife test 0.942		AOPs-SVM
	Non-classical secreted proteins(NCSPs)	ASPIRER: a new computational approach for identifying non-classical secreted proteins based on deep learning	2021	Download ↓ Training dataset: P-141/ N-446; Testing dataset: P-34/ N-34;	SMOTE	5-CV 0.896±0.019	Independent testing 0.8088	ASPIRER
		NonClasGP-Pred: robust and efficient prediction of non-classically secreted proteins by integrating subset-specific optimal models of imbalanced data	2020		/	10-CV 93.23%	Independent testing 86.76%	NonClasGP-Pred
		PeNGaRoo, a combined gradient boosting and ensemble learning framework for predicting non-classical secreted proteins	2020		/	5-CV 0.900±0.016	Independent testing 0.779	PeNGaRo
	Type IV secreted effector proteins	T4SEfinder: a bioinformatics tool for genome-scale prediction of bacterial type IV secreted effectors using pre-trained protein language model	2022	Download ↓ Training data: P-518/ N-1584; Testing dataset: P-20/ N-150;	/	5-CV 90.4±1.4%	Independent testing 96.5%	T4SEfinder
		iT4SE-EP: Accurate Identification of Bacterial Type IV Secreted Effectors by Exploring Evolutionary Features from Two PSI-BLAST Profiles	2021	Download ↓ T4SE_train1502: P-390/ N-1112; T4SE_test180: P-30/ N-150; Train915: P-305/ N-610; Test850: P-75/ N-775;	/	jackknife test Train915: 0.924; Train1502 0.950;	Independent testing Test850: 0.956 Test180: 0.966	iT4SE-EP
		Systematic analysis and prediction of type IV secreted effector proteins by machine learning approaches	2019	Download ↓ Training dataset: P-390/ N-1112; Testing dataset: P-30/ N-150;	/	5-CV 0.957±0.025	Independent testing 0.953±0.014	Bastion4
	Bioluminescent Proteins	Bioluminescent Proteins Prediction with Voting Strategy	2021	Download ↓ BLP_General: P-4604/ N-7093; BLP_Archaea: P-66/ N-748; BLP_Bacteria: P-4362/ N-4919; BLP_Eukaryota: P-176/ N-1426; (ps: 70% training dataset, 30% testing dataset)	/	10-CV 0.939	Independent testing 0.934	/
	Bioluminescent Proteins	iBLP: An XGBoost-Based Predictor for Identifying Bioluminescent Proteins	2021	Download ↓ BLP_General: P-7956/ N-7093; BLP_Archaea: P-45/ N-748; BLP_Bacteria: P-748/ N-4919; BLP_Eukaryota: P-70/ N-1426; (ps: 70% training dataset, 30% testing dataset)	/	10-CV 0.850	Independent testing 0.884	iBLP
	Cell wall lytic enzymes	Predicting Cell Wall Lytic Enzymes Using Combined Features	2021	Download ↓ Lyases: 68 Non-Lyases: 307	SMOTE	jackknife test 99.19%		/
	Electron Transport Proteins	Use Chou’s 5-steps rule with different word embedding types to boost performance of electron transport protein prediction model	2020	Download ↓ Electron transport proteins: 1299 General transport proteins: 4559	/	5-CV 98.5%	Independent testing 96.82%	FastET
	Electron Transport Proteins	Incorporating Distance-Based Top-n-gram and Random Forest To Identify Electron Transport Proteins	2019	Download ↓ Electron transport proteins: 2678 Non-Electron transport proteins: 9630	/	10-CV 84.0%	Independent testing 86.9%	/
	RNA-Binding proteins	rBPDL: Predicting RNA-binding proteins using deep learning	2021	Large Data Download ↓ RBPs sequences: 72226 Negative sequences: 137003	/	10-CV Macro_AUC: 0.932; Micro_AUC: 0.966		rBPDL
		RBPro-RF: Use Chou’s 5-steps rule to predict RNA-binding proteins via random forest with elastic net	2020	Download ↓ Training dataset: RBPs: 2780; Non-RBPs: 7093 Testing dataset: Human: RBPs: 967; Non-RBPs: 597; S. cerevisiae RBPs: 354; Non-RBPs: 135; A. thaliana RBPs: 456; Non-RBPs: 37;	SMOTE	10-CV 97.43%	Independent testing Human: 95.63%; S. cerevisiae: 88.82%; A. thaliana: 92.35%;	RBPro-RF
		TriPepSVM: de novo prediction of RNA-binding proteins based on short amino acid motifs	2019	Download ↓ Training dataset: Huamn(9606): RBPs: 1625; Non-RBPs: 10834; Salmonella(590): RBPs: 275; Non-RBPs: 1273; E.Coli(561): RBPs: 460; Non-RBPs: 3404; Testing dataset: Huamn(9606): RBPs: 181; Non-RBPs: 1204; Salmonella(590): RBPs: 31; Non-RBPs: 142; E.Coli(561): RBPs: 52; Non-RBPs: 379;	/	10-CV Human(9606): AUC 0.83; Salmonella(590): AUC 0.86; E.Coli(561): AUC 0.92;	/	TriPepSVM
	Plant pentatricopeptide repeat	Using a low correlation high orthogonality feature set and machine learning methods to identify plant pentatricopeptide repeat coding gene/protein	2021	Download ↓ Training dataset: P-487/ N-9590	/	10-CV Training set AUC: 0.966; F1: 0.974		/
	Plant pentatricopeptide repeat	Identifying Plant Pentatricopeptide Repeat Coding Gene/Protein Using Mixed Feature Extraction Methods	2019	Download ↓ Training dataset: P-487/ N-9590	/	10-CV Training set AUC: 0.9848; F-measure: 0.9554		MixedPPR
	Sub-Golgi protein	Identification of sub-Golgi protein localization by use of deep representation learning features	2020	Download ↓ D3-Training dataset: cis-Golgi proteins: 101; trans-Golgi proteins: 217; D5-Training dataset: cis-Golgi proteins: 135; trans-Golgi proteins: 1063; D4-Testing dataset: cis-Golgi proteins: 19; trans-Golgi proteins: 51;	SMOTE	LOO D3 92.6% D5 99.2%	Independent testing D3-model: 98.4% D5-model: 96.42%	isGP-DRLF
	Type III secretion systems	EP3: an ensemble predictor that accurately identifies type III secreted effectors	2020	Download ↓ Training dataset1: P-283/ N-313; Training dataset2: P-379/ N-1112; Independent dataset 1: P-35/ N-86; Independent dataset 2: P-83/ N-14; Independent dataset 3: P-108/ N-108; Independent dataset 4: P-226/ N-913;	SMOTE	Independent testing EP3_1_model: Dataset 1: 0.967; Dataset 2: 0.887; Dataset 3: 0.773; Dataset 4: 0.895; EP3_2_model: Dataset 1: 0.818; Dataset 2: 0.629; Dataset 3: 0.922; Dataset 4: 0.838		EP3
	Secretory Proteins	SecProMTB: Support Vector Machine-Based Classifier for Secretory Proteins Using Imbalanced Data Sets Applied to Mycobacterium tuberculosis	2019	Download ↓ Secretory proteins: 35 Non-secretory protein: 266	SMOTE	5-CV 91.6%	Independent testing 86.00%	SecProMTB
Modification Equal length sequence	Multiple PTMs	MusiteDeep: a deep-learning based webserver for protein post-translational modification site prediction and visualization	2020	Large Data Download ↓ Training dataset: Phosphoserine/threonine: P-135556/ N-2803647; Phosphortyrosine: P-9427/ N-93291; N-linked glycosylation: P-90344/ N-511755; O-lined glycosylation: P-4216/ N-103771;N6-acetyllysine: P-22355/ N-274668; Methylarginine: P-4675/ N-99946; Methyllysine: P-2781/ N-45524; S-palmitoylation-cysteine: P-3812/ N-26573; Pyrrolidone-carboxylic-acid: P-1394/ N-10528; Ubiquitination: P-3707/ N-49963; SUMOylation: P-1225/ N-23932; Hydroxylysine: P-356/ N-2650; Hydroxyproline: P-2773/ N-11761; Testing dataset: Phosphoserine/threonine: P-8759/ N-230755; Phosphortyrosine: P-499/ N-5540; N-linked glycosylation: P-20522/ N-120384; O-lined glycosylation: P-218/ N-16248;N6-acetyllysine: P-683/ N-11371; Methylarginine: P-269/ N-6859; Methyllysine: P-154/ N-2001; S-palmitoylation-cysteine: P-151/ N-684; Pyrrolidone-carboxylic-acid: P-230/ N-8918; Ubiquitination: P-514/ N-6621; SUMOylation: P-65/ N-1310; Hydroxylysine: P-9/ N-37; Hydroxyproline: P-422/ N-814;	/	10-CV NA	Independent testing Phosphoserine/threonine: AUC 0.896; Phosphotyrosine: AUC 0.958; N-linked glycosylation: AUC 0.993; O-lined glycosylation: AUC 0.943; N6-acetyllysine: AUC 0.978; Methylarginine: AUC 0.941; Methyllysine: AUC 0.951; S-palmitoylation-cysteine: AUC 0.961; Pyrrolidone-carboxylic-acid: AUC 0.979; Ubiquitination: AUC 0.804; SUMOylation: AUC 0.990; Hydroxylysine: AUC 0.982; Hydroxyproline: AUC 0.732;	MusiteDeep
	Acetylation	Identifying Acetylation Protein by Fusing Its PseAAC and Functional Domain Annotation Running Title: Identifying Acetylation Protein	2019	Download ↓ Acetylation sites samples: 725 Non-acetylation sites samples: 2715	/	5-CV 77.10%		iAcetyp
	Acetylation	Prediction and functional analysis of prokaryote lysine acetylation site by incorporating six types of features into Chou’s general PseAAC	2019	Large Data Download ↓ E.coli: Training dataset: P-6592/ N-15060; Testing dataset: P-361/ N-1384; C.glutamicum: Training dataset: P-1052/ N-6129; Testing dataset: P-83/ N-830; M.tuberculosis: Training dataset: P-865/ N-5167; Testing dataset: P-68/ N-576; B.subtilis: Training dataset: P-1571/ N-12173; Testing dataset: P-125/ N-1165; S.typhimurium: Training dataset: P-198/ N-2477; Testing dataset: P-10/ N-217; G.kaustophilus: Training dataset: P-206/ N-1812; Testing dataset: P-17/ N-192;	/	10-CV E.coli: 0.772; C.glutamicum: 0.756; M.tuberculosis: 0.783; B.subtilis: 0.719; S.typhimurium: 0.821; G.kaustophilus: 0.807;	Independent testing E.coli: 0.851; C.glutamicum: 0.793; M.tuberculosis: 0.827; B.subtilis: 0.831; S.typhimurium: 0.795; G.kaustophilus: 0.809;	PAPred^∗
	Carbonylation	CarSite‑II: an integrated classification algorithm for identifying carbonylated sites based on K‑means similarity‑based undersampling and synthetic minority oversampling techniques	2021	Large Data Download ↓ K Training dataset: P-618/ N-26995; Testing dataset: P-117/ N-7439; P Training dataset: P-162/ N-22418; Testing dataset: P-16/ N-5318; R Training dataset: P-204/ N-22849; Testing dataset: P-54/ N-5966; T Training dataset: P-191 /N-24271; Testing dataset: P-24/ N-6507;	SMOTE-KSU	10-CV K: 82.73%; P: 82.72%; R: 83.16%; T: 85.37%;	Independent testing K: 98.21%; P: 97.92%; R: 97.96%; T: 98.94%;	CarSite‑II
		iCarPS: a computational tool for identifying protein carbonylation sites by novel encoded feature	2021	Download ↓ K Training dataset: P-266/ N-1802; Testing dataset: P-34/ N-147; P Training dataset: P-114/ N-716; Testing dataset: P-12/ N-76; R Training dataset: P-119/ N-754; Testing dataset: P-17/ N-93; T Training dataset: P-116 /N-702; Testing dataset: P-5/ N-30;	/	10-CV K: AUC 0.789; P: AUC 0.814; R AUC 0.726; T: AUC 0.790;	Independent testing K: AUC 0.756; P: AUC 0.752; R: AUC 0.6495; T: AUC 0.8400;	iCarPS
		iCar-PseCp: identify carbonylation sites in proteins by Monte Carlo sampling and incorporating sequence coupled effects into general PseAAC	2020	Download ↓ K P-300/ N-1949; P P-126/ N-792; R P-136/ N-847; T P-121/ N-732;	/	10-CV K: 84.43%; P: 86.79%; R: 84.23%; T: 86.17%;		iCar-PseCp
	Glutarylation	iGlu_AdaBoost: Identification of Lysine Glutarylation Using the AdaBoost Classifier	2021	Download ↓ Training dataset: P-400/ N-1703 Testing dataset: P-44/ N-203	SMOTE-Tomek	10-CV 79.98%	Independent testing 72.07%	/
	Glycosylation	Incorporating a transfer learning technique with amino acid embeddings to efficiently predict N-linked glycosylation sites in ion channels	2021	Download ↓ Training dataset: N-linked glycosylation site:495; Non-N-linked glycosylation site: 5018; Testing dataset: N-linked glycosylation site:103; Non-N-linked glycosylation site: 1019;	/	5-CV 93.4%	Independent testing 92.9%	NIonPred
	Glycosylation	N-GlyDE: a two-stage N-linked glycosylation site prediction incorporating gapped dipeptides and pattern-based encoding	2019	Download ↓ First_stage Training dataset: P-629/ N-5566; Second-stage Training dataset: P-2050/ N-1030 Testing dataset: P-167/ N-280;	/	10-CV 0.733	Independent testing 0.740	N-GlyDE
	Hydroxylation	iHyd-PseCp: Identify hydroxyproline and hydroxylysine in proteins by incorporating sequence-coupled effects into general PseAAC	2016	Download ↓ HyP sites Hydroxylysine sites: 851 Non-Hydroxylysine sites: 3505 HyL sites Hydroxylysine sites: 142 Non-Hydroxylysine sites: 980	/	jackknife test HyP sites: 96.58% HyL sites: 97.08%		iHyd-PseCp
	Malonylation	Incorporating hybrid models into lysine malonylation sites prediction on mammalian and plant proteins	2020	Download ↓ Mammalian Training dataset: P-5006/ N-76264; Testing dataset: P-1252/ N-19066; Plant Training dataset: P-196/ N-2394; Testing dataset: P-82/ N-1195;	/	10-CV Mammalian: 0.764 Plant: 0.660	Independent testing Mammalian: 0.866 Plant: 0.691	Kmalo
	Methylation	DeepRMethylSite: A Deep Learning based approach for Prediction of Arginine Methylation sites in Proteins	2020	Download ↓ Training dataset: P-8344/ N-244600 Testing dataset: P-2085/ N-61150	Undersampling	5-CV 0.76	Independent testing 0.75	DeepRMethylSite
	Palmitoylation	GPS-Palm: a deep learning-based graphic presentation system for the prediction of S-palmitoylation sites in proteins	2020	Download ↓ S-palmitoylation sites: 3089 Non-S-palmitoylation sites: 18992	/	10-CV Human: AUC 0.900 Mouse: AUC 0.897		GPS-Palm
	Phosphorylation	DeepPSP: A Global−Local Information-Based Deep Neural Network for the Prediction of Protein Phosphorylation Sites	2020	Large Data Download ↓ Training dataset: S/T: P-165787/ N-879507; Y: P-28965/ N-134997; Testing dataset: S/T: P-18588/ N-102113; Y: P-3248/ N-14504;	/	S/T: AUC 0.82; Y: AUC 0.73;	NA	DeepPSP
	Pupylation	Recognition of Protein Pupylation Sites by Adopting Resampling Approach	2018	Download ↓ Training dataset: Pupylated sites: 183; Non-pupylated lysine: 2258; Testing dataset: Pupylated lysine: 29; Non-pupylated lysine: 408;	/	10-CV 95.09%	Independent testing 83.75%	/
	S-nitrosylation	Prediction of S-nitrosylation sites by integrating support vector machines and random forest	2019	Download ↓ Training dataset: SNO sites: 33833; Non-SMO sites: 17165; Testing dataset: SNO sites: 351; Non-SNO sites: 3168;	/	5-CV 0.70	Independent testing 0.752	PreSNO
		Identification of S-nitrosylation sites based on multiple features combination	2019	Download ↓ Training dataset: SNO sites: 731; Non-SMO sites: 810; Testing dataset: SNO sites: 124; Non-SNO sites: 221;	/	5-CV 83.11%	Independent testing 73.17%	/
		DeepNitro: Prediction of Protein Nitration and Nitrosylation Sites by Deep Learning	2018	Download ↓ Training dataset: Tyrosine nitration: P-1210/ N-8043;Tryptophan nitration: P-66/ N-155; S-nitrosylation: P-3409/ N-17453; Testing dataset: Tyrosine nitration: P-189/ N-1182; S-nitrosylation: P-485/ N-4947;	/	10-CV Tyrosine nitration: AUC 0.65; Tryptophan nitration: AUC 0.80; S-nitrosylation: AUC 0.70	Independent testing Tyrosine nitration: AUC 0.6879; Tryptophan nitration: AUC 0.8428; S-nitrosylation: AUC 0.70	DeepNitro
	S-sulfenylation	Using deep neural networks and biological subwords to detect protein S-sulfenylation sites	2021	Download ↓ Training dataset: S-sulfenylation sites: 900; Non-S-sulfenylation sites: 6856; Testing dataset: S-sulfenylation sites: 145; Non-S-sulfenylation sites: 268;	/	10-CV 79.9%	Independent testing 77.09%	Sulf-DNN
	S-sulfenylation	SVM-SulfoSite: A support vector machine based predictor for sulfenylation sites	2018		/	10-CV 89.0%	Independent testing 74.0%	SVM-SulfoSite
	Succinylation	Characterization and Identification of Lysine Succinylation Sites based on Deep Learning Method	2019	Download ↓ Training dataset: Succinylation sites: 3216; Non-Succinylation sites: 16412; Testing dataset: Succinylation sites: 218; Non-Succinylation sites: 2621;	/	10-CV 85.68%	Independent testing 86.79%	CNN-SuccSite^∗
	SUMOylation	mUSP: a high-accuracy map of the in situ crosstalk of ubiquitylation and SUMOylation proteome predicted via the feature enhancement approach	2020	Download ↓ Training dataset: SUMOylation sites: 3363; Non-SUMOylation sites: 123131	/	10-CV AUC 0.8472		mUSP^∗
	Ubiquitination	Prediction of protein ubiquitination sites via multi-view features based on eXtreme gradient boosting classifier	2021	Download ↓ Training dataset: Set1: P-150/ N-150; Set2: P-3419/ N-3419; Set3: P-6118/ N-6118; Set4: P-263/ N-4345;Set5: P-131/ N-639; Set6: P-37/ N-639; Testing dataset: Independent test1: P-92/ N-301; Independent test2: P-176/ N-475; Independent test3: P-96/ N-666;	SMOTE	5-CV Set1: AUC 0.8258; Set2: AUC 0.7592; Set3: AUC 0.7853; Set4: AUC 0.9777; Set5: AUC 0.9782; Set6: AUC 0.9860;	Independent testing Independent test1: 78.09%; Independent test2: 74.19%; Independent test3: 87.13%;	UbiSite-XGBoost
	Ubiquitination	Computational identification of ubiquitination sites in Arabidopsis thaliana using convolutional neural networks	2021	Download ↓ Training dataset: Ubiquitination site: 2043; Non-ubiquitination sites: 6130; Testing dataset: Ubiquitination site: 511; Non-ubiquitination sites: 1533;	/	5-CV 85.38%	Independent testing 85.36%	CNNAthUbi
	The other imbalanced classification data of Protein post-translational modifications: ImbClassi_PTMs

Note: 5-CV：5-fold cross validation; 10-CV: 10-fold cross validation; LOO: leave-one-out cross-validation; P: Positive samples; N: Negative samples;
SMOTE: Synthetic Minority Oversampling Technique;
SMOTE-KSU: K-means similaritybased undersampling and the synthetic minority oversampling technique;
SMOTE-Tomek: Synthetic Minority Over-sampling Technique (SMOTE) and the undersampling method Tomek;

Classification based on Protein sequence