ASEB: A Web Server for KAT-specific Acetylation Site Prediction



python.pngPredict based on sequences and functional features


Note: Predict(S+F) is suitable to screen the substrate proteins.

Method: Predict based on sequences and functional features with Swiss-Prot accession number or protein sequence:

Predefine:
CBP/p300     GCN5/PCAF     TIP60/MYST1/2/3/4     HDAC1/HDAC2/HDAC3     SIRT1
Swiss-Prot accession number (recommended) or protein name:

Sequences:



Validation on known human acetylation sites:

Leave-one-out method was adopted to validate the predicting method based on sequences and functions. Each known acetylated peptide (peptide: the lysine site and its surrounding amino acids) was tested by using the other known acetylated peptides as the positive peptide set. A total of 1000 randomly selected peptides from the background database were also tested by using all known acetylated peptides as the positive peptide set.

Known Sensitivity Background Specificity
Total peptides 319 1000
Significant peptides (Votes ≥ 5) 273 85.6% 173 82.7%

Validation on known acetylation sites with independent data set:

Independent data set (from mouse species) was also predicted to validate the predicting method.
1) The mouse acetylated peptides collected previously were used to find their corresponding homologic human proteins with tools from BioMart.
2) Their corresponding lysine site positions were checked with Muscle(Edgar RC., 2004).
3) The identified proteins were deleted and the others were trimmed to acceptable uniforms.
4) Each homologic acetylated site was tested by using all known human acetylated peptides as the positive peptide set.

Positive Sensitivity Negative Specificity
Total peptides 19 1000
Significant peptides (Votes ≥ 5) 19 100.0% 173 82.7%

Validation on known human acetylation sites:

Leave-one-out method was adopted to validate the predicting method based on sequences and functions. Each known acetylated peptide (peptide: the lysine site and its surrounding amino acids) was tested by using the other known acetylated peptides as the positive peptide set. A total of 1000 randomly selected peptides from the background database were also tested by using all known acetylated peptides as the positive peptide set.

Known Sensitivity Background Specificity
Total peptides 107 1000
Significant peptides (Votes ≥ 5) 71 66.4% 105 89.5%

Validation on known acetylation sites with independent data set:

Independent data set (from mouse species) was also predicted to validate the predicting method.
1) The mouse acetylated peptides collected previously were used to find their corresponding homologic human proteins with tools from BioMart.
2) Their corresponding lysine site positions were checked with Muscle(Edgar RC., 2004).
3) The identified proteins were deleted and the others were trimmed to acceptable uniforms.
4) Each homologic acetylated site was tested by using all known human acetylated peptides as the positive peptide set.

Positive Sensitivity Negative Specificity
Total peptides 7 1000
Significant peptides (Votes ≥ 5) 5 71.4% 105 89.5%

Validation on known human deacetylation sites:

Leave-one-out method was adopted to validate the predicting method based on sequences and functions. Each known deacetylated peptide (peptide: the lysine site and its surrounding amino acids) was tested by using the other known deacetylated peptides as the positive peptide set. A total of 1000 randomly selected peptides from the background database were also tested by using all known deacetylated peptides as the positive peptide set.

Known Sensitivity Background Specificity
Total 40 1000
Significant peptides (Votes ≥ 5) 31 77.5% 43 95.7%

Validation on known human deacetylation sites:

Leave-one-out method was adopted to validate the predicting method based on sequences and functions. Each known deacetylated peptide (peptide: the lysine site and its surrounding amino acids) was tested by using the other known deacetylated peptides as the positive peptide set. A total of 1000 randomly selected peptides from the background database were also tested by using all known deacetylated peptides as the positive peptide set.

Known Sensitivity Background Specificity
Total 118 1000
Significant peptides (Votes ≥ 5) 89 75.4% 108 89.2%

Validation on known deacetylation sites with independent data set:

Independent data set (from mouse species) was also predicted to validate the predicting method.
1) The mouse deacetylated peptides collected previously were used to find their homologic human proteins with tools from BioMart.
2) Their corresponding lysine site positions were checked with Muscle(Edgar RC., 2004).
3) The identified proteins were deleted and the others were trimmed to acceptable uniforms.
4) Each homologic deacetylated site was tested by using all known human deacetylated peptides as the positive peptide set.

Positive Sensitivity Negative Specificity
Total 227 1000
Significant peptides (Votes ≥ 5) 107 47.1% 108 89.2%

Validation with biological experiments:

For potential substrates of SIRT1, verification assays by immunoprecipitation combined with western blotting were conducted. And six of nine predicted substrates for SIRT1 were confirmed.

Testing proteins Sensitivity
Total proteins 9
Significant proteins 6 66.7%

Validation on known human acetylation sites:

Leave-one-out method was adopted to validate the predicting method based on sequences and functions. Each known acetylated peptide (peptide: the lysine site and its surrounding amino acids) was tested by using the other known acetylated peptides as the positive peptide set. A total of 1000 randomly selected peptides from the background database were also tested by using all known acetylated peptides as the positive peptide set.

Known Sensitivity Background Specificity
Total 28 1000
Significant peptides (Votes ≥ 5) 14 50.0% 44 95.6%

Validation with biological experiments:

For potential substrates of MYST family , verification assays by immunoprecipitation combined with western blotting were conducted. And three of six predicted substrates for MYST family were confirmed.

Testing proteins Sensitivity
Total proteins 6
Significant proteins 3 50.0%