Deep-PLA | Help

What is Deep-PLA

Deep-PLA is a web server based on deep learning to predict enzyme-specific acetylation regulation. Given a protein or a set of proteins in FASTA format, it can predict the possible acetylation sites by several HATs and HDACs, including CREBBP, EP300, HAT1, KAT2A, KAT2B, KAT5, KAT8, HDAC1, HDAC2, HDAC3, HDAC6, SIRT1, SIRT2, SIRT3, SIRT6, and SIRT7. Deep-PLA adopted four feature extraction methods to extract the sequence information and used a deep neural network to learn them, in the meanwhile, PSO (particle swarm optimizer) algorithm was used to fine turn the hyperparameters such as learning rate, layer numbers, activation functions and so on. Deep-PLA performed better than previous predictors for HAT/HDAC-specific acetylation.

About predicting performance

ROC for cross validation

ROC for test validation

Cross validation performance

Enzyme	Positive	Negative	Threshold	Specificity	Sensitivity	Precision

Enzyme	Positive	Negative	Threshold	Specificity	Sensitivity	Precision
CREBBP	217	2414	High	0.9519	0.1842	0.2298
			Medium	0.9059	0.2632	0.1786
			Low	0.8633	0.3958	0.2249
EP300	511	4832	High	0.9507	0.2947	0.3691
			Medium	0.9052	0.4495	0.3500
			Low	0.8573	0.5229	0.2938
HAT1	12	145	High	0.9630	0.6000	0.7500
			Medium	0.9552	0.6667	0.7067
			Low	0.9259	0.8000	0.6794
KAT2A	40	446	High	0.9470	0.5714	0.5170
			Medium	0.8981	0.7143	0.3841
			Low	0.8528	0.8571	0.3468
KAT2B	108	1860	High	0.9634	0.1600	0.2308
			Medium	0.8976	0.4211	0.2017
			Low	0.7646	0.4375	0.0769
KAT5	52	784	High	0.9542	0.4667	0.5000
			Medium	0.8856	0.7333	0.3874
			Low	0.6830	0.8667	0.2137
KAT8	23	311	High	0.9367	0.2857	0.4358
			Medium	0.8782	0.5000	0.3326
			Low	0.8475	0.6250	0.3571
HDAC1	47	827	High	0.9649	0.4615	0.5532
			Medium	0.9199	0.5385	0.4092
			Low	0.8652	0.6429	0.3253
HDAC2	28	534	High	0.9515	0.7000	0.5916
			Medium	0.8791	0.7143	0.3375
			Low	0.8126	0.9167	0.4020
HDAC3	37	500	High	0.9542	0.4444	0.5134
			Medium	0.8965	0.5556	0.3878
			Low	0.8053	0.6154	0.3097
HDAC6	17	365	High	0.9348	0.3750	0.4810
			Medium	0.8116	0.6250	0.2784
			Low	0.5463	0.7778	0.1792
HDACs	147	1738	High	0.9405	0.7105	0.6158
			Medium	0.8958	0.7600	0.3815
			Low	0.8220	0.7632	0.3292
SIRT1	185	2982	High	0.9485	0.5143	0.3738
			Medium	0.8947	0.5806	0.2250
			Low	0.8692	0.6136	0.2598
SIRT2	52	733	High	0.9607	0.2353	0.4484
			Medium	0.9235	0.3125	0.3808
			Low	0.8758	0.3571	0.2543
SIRT3	30	522	High	0.9643	0.0769	0.3397
			Medium	0.9082	0.1538	0.1849
			Low	0.8163	0.2308	0.1494
SIRT6	10	130	High	0.9500	0.3750	0.7833
			Medium	0.7223	0.5000	0.2498
			Low	0.6400	0.6667	0.2745
SIRT7	12	279	High	0.8269	0.1429	0.1000
			Medium	0.8173	0.2857	0.1742
			Low	0.7981	0.4286	0.2225

Testing validation performance

Enzyme	Positive	Negative	Threshold	Specificity	Sensitivity	Precision

Enzyme	Positive	Negative	Threshold	Specificity	Sensitivity	Precision
CREBBP	87	631	High	0.9509	0.2529	0.4151
			Medium	0.9002	0.4138	0.3636
			Low	0.8510	0.5057	0.3188
EP300	104	876	High	0.9509	0.2885	0.4110
			Medium	0.9007	0.4135	0.3308
			Low	0.8505	0.5673	0.3105
HAT1	5	33	High	0.9697	1.0000	0.8333
			Medium	0.9091	1.0000	0.6250
			Low	0.8485	1.0000	0.5000
KAT2A	40	466	High	0.9506	0.8250	0.5893
			Medium	0.9013	0.9500	0.4524
			Low	0.8519	0.9500	0.3551
KAT2B	43	192	High	0.9531	0.0930	0.3077
			Medium	0.9010	0.1628	0.2692
			Low	0.8542	0.3953	0.3778
KAT5	6	47	High	0.9574	1.0000	0.7500
			Medium	0.9149	1.0000	0.6000
			Low	0.8511	1.0000	0.4615
KAT8	6	174	High	0.9540	0.3333	0.2000
			Medium	0.9023	0.3333	0.1053
			Low	0.8506	0.3333	0.0714
HDAC1	4	61	High	0.9508	0.5000	0.4000
			Medium	0.9016	0.5000	0.2500
			Low	0.8525	0.5000	0.1818
HDAC2	1	20	High	0.7500	1.0000	0.1667
			Medium	0.7000	1.0000	0.1429
			Low	0.6500	1.0000	0.1250
HDAC3	2	70	High	0.9571	0.5000	0.2500
			Medium	0.9000	1.0000	0.2222
			Low	0.8571	1.0000	0.1667
HDAC6	16	316	High	0.9525	0.2500	0.2105
			Medium	0.9019	0.4375	0.1842
			Low	0.8513	0.5000	0.1455
HDACs	26	549	High	0.9508	0.2692	0.2059
			Medium	0.9016	0.3462	0.1429
			Low	0.8506	0.4615	0.1277
SIRT1	33	601	High	0.9501	0.3636	0.2857
			Medium	0.9002	0.3939	0.1781
			Low	0.8502	0.4242	0.1346
SIRT2	9	145	High	0.9517	0.4444	0.3636
			Medium	0.9034	0.4444	0.2222
			Low	0.8552	0.6667	0.2222
SIRT3	17	190	High	0.9526	0.3529	0.4000
			Medium	0.9000	0.3529	0.2400
			Low	0.8526	0.5294	0.2432
SIRT6	2	11	High	0.9091	0.5000	0.5000
			Medium	0.8182	0.5000	0.3333
			Low	0.7273	0.5000	0.2500
SIRT7	4	25	High	0.9600	0.5000	0.6667
			Medium	0.9200	0.5000	0.5000
			Low	0.8400	0.5000	0.3333

Usage procedures

Step1: Paste your FASTA format sequence into the input textarea, or you can click the example button to run the default sequence.

Step2: Choose a kind of enzyme, if ‘All’ is selected, all enzymes will be predicted.

Step3: Choose a threshold to get a high confidence result, the default value is ‘Medium’.

Step4: Click Submit.

Output result explanation

Potential HAT(s)/HDAC(s) modification site(s)

This table contains five column, including FASTA title ID, position, enzyme, FPR (false positive rate) and the peptide sequence window.

Network and colocalization

The PPI data are collected from several databases, the blue nodes are the HATs, while the orange nodes are the HDACs, the red node is the protein that matches query sequence best and green nodes are scaffolds.

Secondary structure and surface accessibility

The disorder values are calculated by IUPred. The surface accessibility and secondary structure information are predicted by NetSurfP. By the way, the predicted acetylation and deacetylation sites are labeled.

Navigation