Table 2 Human proteins with the highest predicted relevances in ageing.

From: Prediction and characterization of human ageing-related proteins by using machine learning

Uniprot ID

recommended name in UniProt

ageing neighbours

“aging” GO

GenAge

average predicted value

BCL2_HUMAN

Apoptosis regulator Bcl-2

4

yes

yes

0.981

FOXO1_HUMAN

Forkhead box protein O1

4

no

yes

0.96

ERCC1_HUMAN

DNA excision repair protein ERCC-1

3

yes

yes

0.944

PCNA_HUMAN

Proliferating cell nuclear antigen

4

no

yes

0.936

FOXO3_HUMAN

Forkhead box protein O3 {ECO:0000305}

5

yes

yes

0.929

SIR2_HUMAN

NAD-dependent protein deacetylase sirtuin-2

2

no

no

0.909

PTEN_HUMAN

Phosphatidylinositol 3,4,5-trisphosphate 3-phosphatase and dual-specificity protein phosphatase

5

yes

yes

0.882

APEX1_HUMAN

DNA-(apurinic or apyrimidinic site) lyase

2

yes

yes

0.857

HDAC2_HUMAN

Histone deacetylase 2

3

no

yes

0.849

MTOR_HUMAN

Serine/threonine-protein kinase mTOR

3

yes

yes

0.832

BECN1_HUMAN

Beclin-1

3

yes

no

0.827

AKT1_HUMAN

RAC-alpha serine/threonine-protein kinase

10

yes

yes

0.827

KPCD_HUMAN

Protein kinase C delta type

3

yes

yes

0.808

CDK1_HUMAN

Cyclin-dependent kinase 1

2

yes

yes

0.804

SYUA_HUMAN

Alpha-synuclein

2

yes

no

0.801

P73_HUMAN

Tumor protein p73

2

no

yes

0.8

PARP1_HUMAN

Poly [ADP-ribose] polymerase 1

6

no

yes

0.798

PRKDC_HUMAN

DNA-dependent protein kinase catalytic subunit

4

no

yes

0.791

ABL1_HUMAN

Tyrosine-protein kinase ABL1

6

no

yes

0.782

WRN_HUMAN

Werner syndrome ATP-dependent helicase

9

yes

yes

0.782

  1. The 20 highest scored proteins considered the entire set of human proteins (regardless of whether or not the protein is included in the GenAge database), sorted by decreasing predicted relevance in ageing (average predicted value). Each row consists of an ID of the given protein (“Uniprot ID”), a description (“recommended name in UniProt”), the number of ageing-related protein neighbours of the given protein in the protein-protein interaction network (“ageing neighbours”), a statement about its assignment to the GO term “aging” (“aging GO”), a statement about its inclusion in GenAge (“GenAge”), and the average predicted value of 20 predictions of three machine learning methods each (XGBoost, SVM and LR) by using the final feature set selected by XGBoost (“average predicted value”). Average predicted values close to one indicate very strong predicted relevance for the human ageing process. Supplementary Table S1 is a more detailed list with all of the human proteins.
Back to article page