Exploratory data analysis

Setup

import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
import seaborn as sns
from matplotlib.colors import LogNorm
from katlas.core import *

sns.set(rc={"figure.dpi":300, 'savefig.dpi':300})
sns.set_context('notebook')
sns.set_style("ticks")

Overlap between sources

# direct concatenate of all source sequence
df=pd.read_parquet('raw/combine_source.parquet')

# unique substrate site with source split by | 
df_group=pd.read_parquet('raw/combine_source_grouped.parquet')

def plot_overlap(df_concat,source_col='source',id_col='kin_sub_site',figsize=(7,5),title="Overlap of Unique KS Pairs Between Sources"):
    source_ids = df_concat.groupby(source_col)[id_col].apply(set).to_dict()
    sources = list(source_ids.keys())
    
    overlap_matrix = pd.DataFrame(index=sources, columns=sources, dtype=int)
    for src1 in sources:
        for src2 in sources:
            if src1 == src2:
                overlap_matrix.loc[src1, src2] = len(source_ids[src1])
            else:
                overlap_matrix.loc[src1, src2] = len(source_ids[src1].intersection(source_ids[src2]))
    
    mask = np.triu(np.ones_like(overlap_matrix, dtype=bool), k=1)

    overlap_matrix = overlap_matrix.astype(int)
    
    vmin = np.min(overlap_matrix[overlap_matrix > 0])  # Smallest nonzero value
    vmax = np.max(overlap_matrix)
    
    plt.figure(figsize=figsize)
    sns.heatmap(overlap_matrix, 
                annot=True, 
                mask=mask, 
                fmt=",", 
                # vmin=1, vmax=1,
                norm=LogNorm(vmin=vmin, vmax=vmax),
                cmap="Blues", 
                cbar=False, 
                linewidths=1, 
                linecolor='white')

    plt.title(title)

plot_overlap(df)

df.site.str[0].value_counts()

site
Y    92088
S    84341
T    35243
Name: count, dtype: int64

df_st = df[df.site.str[0]!='Y']
df_y = df[df.site.str[0]=='Y']

plot_overlap(df_st,title="Overlap of KS Pairs in S/T sites Between Sources")
plot_overlap(df_y,title="Overlap of KS Pairs in Y sites Between Sources")

Overlap of phosphosites with human phosphoproteome

# data = Data.get_combine_site_phosphorylated()

df=pd.read_parquet('raw/combine_source.parquet')
data = pd.read_parquet('raw/human_phosphoproteome.parquet')

data = data.rename(columns={'uniprot':'substrate_uniprot'})

def get_key(df):
    df = df.copy()
    print('original shape:',df.shape)
    df['sub_site']=df['substrate_uniprot']+'_'+df['site']
    df = df.drop_duplicates(subset='sub_site')
    print('after removing duplicates',df.shape)
    return df

data_site = get_key(data)
df_site=get_key(df)

original shape: (119955, 11)
after removing duplicates (119955, 11)
original shape: (211672, 7)
after removing duplicates (29811, 8)

data_site = data_site[['sub_site']].copy()
df_site = df_site[['sub_site']].copy()

data_site['source']='Human Phosphoprotome'
df_site['source']='KS Datasets'

df_concat = pd.concat([data_site,df_site],ignore_index=True)

df_concat_group = df_concat.groupby('sub_site')['source'].agg(lambda x: '|'.join(x)).reset_index()

df_concat_group.source.value_counts()

source
Human Phosphoprotome                102032
Human Phosphoprotome|KS Datasets     17923
KS Datasets                          11888
Name: count, dtype: int64

# plot_overlap(df_concat,source_col='source',id_col='sub_site',figsize=(4,2),
#              title="Overlap of Unique Substrate Sites Between Datasets")

Venn diagram:

# pip install matplotlib_venn

from matplotlib_venn import venn2

set_human = set(data_site['sub_site'])
set_ks = set(df_site['sub_site'])

plt.figure(figsize=(5,5))
venn = venn2([set_human, set_ks], 
             # set_labels=('Human Phosphoproteome', 'KS Datasets')
             set_labels=('', '')
            )

for label in venn.subset_labels:
    if label:  # Check if the label exists (not empty subset)
        label.set_text(f"{int(label.get_text()):,}")  # Add comma formatting
plt.gca().set_aspect(0.8)
plt.title("Overlap of Unique Substrate Sites Between Datasets")
plt.show()

Stacked plot of S/T/Y compositions

df['acceptor']=df['site'].str[0]

df_sugi = df[df.source=='Sugiyama'].reset_index(drop=True)
df_rest = df[df.source!='Sugiyama'].reset_index(drop=True)

import matplotlib.ticker as mticker

def plot_stacked(df,figsize=(6,4)):
    plt.figure(figsize=figsize)
    sns.histplot(data=df, x='source', hue='acceptor', multiple='stack', discrete=True, shrink=0.8)
    
    plt.xlabel('')
    plt.ylabel('Number of Sites')
    plt.title('Total Number of Unique KS Pairs per Source')
    plt.gca().yaxis.set_major_formatter(mticker.FuncFormatter(lambda x, _: f'{int(x):,}')) # show number with comma
    plt.xticks(rotation=45)

plot_stacked(df_rest)

def plot_stacked2(df,figsize=(6,4)):
    plt.figure(figsize=figsize)
    sns.histplot(data=df, x='source', hue='acceptor', multiple='stack', discrete=True, shrink=0.8)
    
    plt.xlabel('')
    plt.ylabel('')
    plt.xticks(rotation=45)
    plt.legend([],frameon=False)
    plt.gca().yaxis.set_major_formatter(mticker.FuncFormatter(lambda x, _: f'{int(x):,}'))
    plt.tick_params(axis='y', labelright=True, labelleft=False, right=True, left=False)

Uncomment below to plot the sugiyama site number:

# plot_stacked2(df_sugi,(0.7,4))

Recovery rate

Check raw/recovery_rate.xlsx for more info.

The data is collected by: - Filter only human first, get dataframe shape - After kinase & substrate mapping, get the shape - After sites validation, get the shape

Co-occured kinase

df_group['substrate_uniprot_site']=df_group['substrate_uniprot']+'_'+df_group['site']

from itertools import combinations
from collections import Counter

df_group.groupby('substrate_uniprot_site')['kinase_uniprot'].apply(list)

substrate_uniprot_site
A0A2R8Y4L2_S158                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      [P48730]
A0A2R8Y4L2_S22                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       [O14757, O95819, Q13557]
A0A2R8Y4L2_S6                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        [O15264, P06493, Q9NR20]
A0A2R8Y4L2_S95     [O00238, O14920, O15075, O43318, O75676, O94806, O96017, P05771-2, P05771, P15735, P17252, P23443, P24723, P24941, P36507, P36896, P36897, P41743, P42679, P49137, P49759, P49761, P50750, P51812, P51955, P51957, P52333, P57059, P78362, P80192, Q02156, Q02779, Q04759, Q04771, Q05513, Q06187, Q13131, Q13554, Q13555, Q13557, Q14012, Q14680, Q15139, Q15349, Q16512, Q16513, Q16566, Q16584, Q16644, Q16816, Q86UE8, Q8IU85, Q8IW41, Q8N568, Q8TD19, Q8TDX7, Q96GD4, Q9BYT3, Q9BZL6, Q9NYL2, Q9UKI8, Q9UPE1, Q9UQM7, Q9Y2U5, Q9Y4K4]
A0A2R8Y4L2_T25                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       [O14757, O95819, Q13557]
                                                                                                                                                                                                                                                                                      ...                                                                                                                                                                                                                                                                    
Q9Y6Y8_S894                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  [Q96PF2, Q9BXA7]
Q9Y6Y8_T893                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  [P27448, Q7KZI7, Q96L34, Q9P0L2]
Q9Y6Y8_Y935                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  [P43405, P52333]
Q9Y6Y9_Y131                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          [P07948]
Q9Y6Y9_Y22                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           [P07948]
Name: kinase_uniprot, Length: 29811, dtype: object

# Count co-occurring kinase pairs
co_occurrence = Counter()
for kinases in grouped:
    for pair in combinations(sorted(set(kinases)), 2):  # remove duplicates and sort for consistency
        co_occurrence[pair] += 1

# print(co_occurrence)

kinases = set()
for pair in co_occurrence.keys():
    kinases.update(pair)
kinases = sorted(kinases)

# Initialize the co-occurrence matrix with zeros
matrix = pd.DataFrame(0, index=kinases, columns=kinases)

# Fill in the counts (matrix is symmetric)
for (k1, k2), count in co_occurrence.items():
    matrix.loc[k1, k2] = count
    matrix.loc[k2, k1] = count

matrix['P00533'].sort_values()

P31152      0
Q16659      0
Q16654      0
P37173      0
P35790      0
         ... 
P36888    478
P11362    480
P07949    480
P06213    481
Q15303    497
Name: P00533, Length: 433, dtype: int64

counts = df_group.groupby('kinase_uniprot').agg({'substrate_uniprot_site':'size'})

Jaccard

counts_series = counts['substrate_uniprot_site']

# Create a copy for normalized values
norm_matrix = matrix.copy().astype(float)

for k1 in matrix.index:
    for k2 in matrix.columns:
        intersection = matrix.loc[k1, k2]
        union = counts_series[k1] + counts_series[k2] - intersection
        norm_matrix.loc[k1, k2] = intersection / union if union != 0 else 0
    norm_matrix.loc[k1, k1] = 1 # diaganol to be 1

# Create a boolean mask for the upper triangle (k=1 excludes the diagonal)
mask = np.triu(np.ones(norm_matrix.shape), k=1).astype(bool)
norm_matrix = norm_matrix.where(mask)

matrix['P00533'].sort_values(ascending=False).head()

Q15303    497
P06213    481
P07949    480
P11362    480
P36888    478
Name: P00533, dtype: int64

norm_matrix['P00533'].sort_values(ascending=False).head()

O60674    0.253078
O15146    0.213529
P00519    0.203051
O43318    0.122415
O75676    0.023234
Name: P00533, dtype: float64

kin_info = Data.get_kinase_info()

st=kin_info[kin_info.group!='TK'].uniprot.tolist()

st_col =norm_matrix.columns[norm_matrix.columns.isin(st)]

st_matrix = norm_matrix[st_col]

def get_max_jaccard(matrix):   
    a = matrix.max()
    b = matrix.idxmax()
    
    matrix = pd.concat([a, b], axis=1)
    matrix.columns = ['jaccard', 'kinase_uniprot']
    
    return matrix.sort_values('jaccard', ascending=False)

get_max_jaccard(st_matrix)

/tmp/ipykernel_169440/3714721923.py:3: FutureWarning: The behavior of DataFrame.idxmax with all-NA values, or any-NA and skipna=False, is deprecated. In a future version this will raise ValueError
  b = matrix.idxmax()

	jaccard	kinase_uniprot
Q9NYV4	0.509804	Q14004
Q96QP1	0.500000	Q96L96
Q9Y6E0	0.465691	Q9P289
P28482	0.450737	P27361
P31751	0.437574	P31749
...	...	...
O76039	0.000000	O00141
Q2M2I8	0.000000	O00141
O60885	0.000000	O00141
O00311	0.000000	O00141
O00141	NaN	NaN

323 rows × 2 columns

tyr=kin_info[kin_info.group=='TK'].uniprot.tolist()
tyr_col =norm_matrix.columns[norm_matrix.columns.isin(tyr)]
tyr_matrix = norm_matrix[tyr_col]

get_max_jaccard(tyr_matrix)

	jaccard	kinase_uniprot
P54762	0.554638	P54760
Q15375	0.539118	P21709
P54753	0.531496	P29320
Q06418	0.526057	P54762
P54760	0.521168	P54753
...	...	...
P04626	0.046957	O43318
O15146	0.044005	O00506
Q6J9G0	0.016393	O43293
Q6ZMQ8	0.004348	O96017
Q8IWU2	0.003676	Q13976

83 rows × 2 columns

kin_info_uniprot = pd.read_excel('raw/uniprot_human_keyword_kinase.xlsx')

kin_map = kin_info_uniprot.set_index('Entry')['Entry Name'].str.split('_').str[0]

names = norm_matrix.index.map(kin_map)

todo: Use filter to only select kinase >30 counts

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
from scipy.cluster.hierarchy import linkage, dendrogram
from sklearn.manifold import MDS
import networkx as nx

# Assuming norm_matrix is your pre-computed Jaccard similarity matrix

# ---------------------------
# 1. Clustering & Heatmap

# Alternatively, using hierarchical clustering with scipy:
# Z = linkage(norm_matrix, method='average')
# plt.figure(figsize=(10, 5))
# dendrogram(Z, labels=norm_matrix.index)
# plt.title("Hierarchical Clustering Dendrogram")
# plt.xlabel("Kinase")
# plt.ylabel("Distance")
# plt.show()

Z = linkage(norm_matrix, method='average')
plt.figure(figsize=(4,50))
dendrogram(Z,orientation='left',leaf_font_size=7,labels=names,color_threshold=1.3)
plt.title('Hierarchical Clustering Dendrogram')
plt.ylabel('Distance')
plt.show()

# 2. Network Analysis
# ---------------------------
# Build a graph from norm_matrix (using a threshold to filter weak links)
threshold = 0.2  # adjust threshold as needed
G = nx.Graph()
for i, kinase1 in enumerate(norm_matrix.index):
    for j, kinase2 in enumerate(norm_matrix.columns):
        if i < j and norm_matrix.loc[kinase1, kinase2] > threshold:
            G.add_edge(kinase1, kinase2, weight=norm_matrix.loc[kinase1, kinase2])

# Draw the network
plt.figure(figsize=(10, 10))
pos = nx.spring_layout(G, weight='weight')
nx.draw(G, pos, with_labels=True, node_size=500, font_size=8)
plt.title("Kinase Similarity Network")
plt.show()

Frequency of kinase

Instead of showing PSSM, we can show the neigboring kinase

Kinase count

df_group['substrate_uniprot_site']=df_group['substrate_uniprot']+df_group['site']

counts = df_group.groupby('kinase_uniprot').agg({'substrate_uniprot_site':'size'}).reset_index()

counts

	kinase_uniprot	substrate_uniprot_site
0	O00141	387
1	O00238	356
2	O00311	50
3	O00329	6
4	O00418	202
...	...	...
450	Q9Y572	10
451	Q9Y5S2	109
452	Q9Y6E0	846
453	Q9Y6M4	224
454	Q9Y6R4	15

455 rows × 2 columns

kin_info = pd.read_excel('raw/uniprot_human_keyword_kinase.xlsx')

kin_info = kin_info.rename(columns={'Entry':'kinase_uniprot'})

kin_info2 = Data.get_kinase_info()

(kin_info2.in_ST_paper==1).sum()

303

(kin_info2.in_Tyr_paper==1).sum()

93

in_pspa=(kin_info2.in_ST_paper==1)|(kin_info2.in_Tyr_paper==1)

kin_info2[in_pspa]

	kinase	ID_coral	uniprot	ID_HGNC	group	family	subfamily_coral	subfamily	in_ST_paper	in_Tyr_paper	...	cytosol	cytoskeleton	plasma membrane	mitochondrion	Golgi apparatus	endoplasmic reticulum	vesicle	centrosome	aggresome	main_location
0	AAK1	AAK1	Q2M2I8	AAK1	Other	NAK	NaN	NAK	1	0	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
1	ABL1	ABL1	P00519	ABL1	TK	Abl	NaN	Abl	0	1	...	6.0	NaN	4.0	NaN	NaN	NaN	NaN	NaN	NaN	cytosol
2	ABL2	ABL2	P42684	ABL2	TK	Abl	NaN	Abl	0	1	...	4.0	6.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	cytoskeleton
3	TNK2	ACK	Q07912	TNK2	TK	Ack	NaN	Ack	0	1	...	NaN	NaN	NaN	NaN	NaN	NaN	8.0	NaN	2.0	vesicle
4	ACVR2A	ACTR2	P27037	ACVR2A	TKL	STKR	STKR2	STKR2	1	0	...	5.0	NaN	NaN	NaN	NaN	5.0	NaN	NaN	NaN	cytosol
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
518	YSK1	YSK1	O00506	STK25	STE	STE20	YSK	YSK	1	0	...	6.0	NaN	NaN	NaN	4.0	NaN	NaN	NaN	NaN	cytosol
519	ZAK	ZAK	Q9NYL2	MAP3K20	TKL	MLK	ZAK	ZAK	1	0	...	5.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	nucleus
520	ZAP70	ZAP70	P43403	ZAP70	TK	Syk	NaN	Syk	0	1	...	5.0	NaN	2.0	NaN	NaN	NaN	NaN	NaN	NaN	cytosol
521	EEF2K	eEF2K	O00418	EEF2K	Atypical	Alpha	eEF2K	eEF2K	1	0	...	9.0	NaN	1.0	NaN	NaN	NaN	NaN	NaN	NaN	cytosol
522	FAM20C	FAM20C	Q8IXL6	FAM20C	Atypical	FAM20C	NaN	FAM20C	1	0	...	2.0	NaN	NaN	NaN	7.0	1.0	NaN	NaN	NaN	Golgi apparatus

392 rows × 30 columns

kin_info2[

Index(['kinase', 'ID_coral', 'uniprot', 'ID_HGNC', 'group', 'family',
       'subfamily_coral', 'subfamily', 'in_ST_paper', 'in_Tyr_paper',
       'in_cddm', 'pseudo', 'pspa_category_small', 'pspa_category_big',
       'cddm_big', 'cddm_small', 'length', 'human_uniprot_sequence',
       'kinasecom_domain', 'nucleus', 'cytosol', 'cytoskeleton',
       'plasma membrane', 'mitochondrion', 'Golgi apparatus',
       'endoplasmic reticulum', 'vesicle', 'centrosome', 'aggresome',
       'main_location'],
      dtype='object')

df_group[df_group.kinase_uniprot=='P07948-2']

	kin_sub_site	kinase_uniprot	substrate_uniprot	site	source	substrate_genes	substrate_sequence	substrate_phosphoseq	substrate_uniprot_site
30312	P07948-2_A0MZ66_Y24	P07948-2	A0MZ66	Y24	Sugiyama	SHTN1 KIAA1598	MNSSDEEKQLQLITSLKEQAIGEYEDLRAENQKTKEKCDKIRQERD...	MNSSDEEKQLQLITSLKEQAIGEyEDLRAENQKTKEKCDKIRQERD...	A0MZ66Y24
30313	P07948-2_A1L429_Y10	P07948-2	A1L429	Y10	Sugiyama	GAGE12B; GAGE12C; GAGE12D; GAGE12E	MSWRGRSTYYWPRPRRYVQPPEMIGPMRPEQFSDEVEPATPEEGEP...	MSWRGRsTyyWPRPRRYVQPPEMIGPMRPEQFSDEVEPATPEEGEP...	A1L429Y10
30314	P07948-2_A1L429_Y9	P07948-2	A1L429	Y9	Sugiyama	GAGE12B; GAGE12C; GAGE12D; GAGE12E	MSWRGRSTYYWPRPRRYVQPPEMIGPMRPEQFSDEVEPATPEEGEP...	MSWRGRsTyyWPRPRRYVQPPEMIGPMRPEQFSDEVEPATPEEGEP...	A1L429Y9
30315	P07948-2_A1X283_Y855	P07948-2	A1X283	Y855	Sugiyama	SH3PXD2B FAD49 KIAA1295 TKS4	MPPRRSIVEVKVLDVQKRRVPNKHYVYIIRVTWSSGSTEAIYRRYS...	MPPRRSIVEVKVLDVQKRRVPNKHYVYIIRVTWSSGSTEAIYRRYS...	A1X283Y855
30316	P07948-2_A5A3E0_S939	P07948-2	A5A3E0	S939	Sugiyama	POTEF A26C1B	MVVEVDSMPAASSVKKPFGLRSKMGKWCCRCFPCCRESGKSNVGTS...	MVVEVDSMPAASSVKKPFGLRSKMGKWCCRCFPCCRESGKSNVGTS...	A5A3E0S939
...	...	...	...	...	...	...	...	...	...
31872	P07948-2_Q9Y5S9_Y54	P07948-2	Q9Y5S9	Y54	Sugiyama	RBM8A RBM8 HSPC114 MDS014	MADVLDLHEAGGEDFAMDEDGDESIHKLKEKAKKRKGRGFGSEEGS...	MADVLDLHEAGGEDFAMDEDGDESIHKLKEKAKKRKGRGFGSEEGS...	Q9Y5S9Y54
31873	P07948-2_Q9Y639_Y216	P07948-2	Q9Y639	Y216	Sugiyama	NPTN SDFR1 SDR1	MSGSSLPSALALSLLLVSGSLLPGPGAAQNAGFVKSPMSETKLTGD...	MSGSSLPSALALSLLLVSGSLLPGPGAAQNAGFVKSPMSETKLTGD...	Q9Y639Y216
31874	P07948-2_Q9Y639_Y220	P07948-2	Q9Y639	Y220	Sugiyama	NPTN SDFR1 SDR1	MSGSSLPSALALSLLLVSGSLLPGPGAAQNAGFVKSPMSETKLTGD...	MSGSSLPSALALSLLLVSGSLLPGPGAAQNAGFVKSPMSETKLTGD...	Q9Y639Y220
31875	P07948-2_Q9Y696_S245	P07948-2	Q9Y696	S245	Sugiyama	CLIC4	MALSMPLNGLKEEDKEPLIELFVKAGSDGESIGNCPFSQRLFMILW...	MALSMPLNGLKEEDKEPLIELFVKAGsDGEsIGNCPFSQRLFMILW...	Q9Y696S245
31876	P07948-2_Q9Y696_Y244	P07948-2	Q9Y696	Y244	Sugiyama	CLIC4	MALSMPLNGLKEEDKEPLIELFVKAGSDGESIGNCPFSQRLFMILW...	MALSMPLNGLKEEDKEPLIELFVKAGsDGEsIGNCPFSQRLFMILW...	Q9Y696Y244

1565 rows × 9 columns

counts

	kinase_uniprot	substrate_uniprot_site
0	O00141	387
1	O00238	356
2	O00311	50
3	O00329	6
4	O00418	202
...	...	...
450	Q9Y572	10
451	Q9Y5S2	109
452	Q9Y6E0	846
453	Q9Y6M4	224
454	Q9Y6R4	15

455 rows × 2 columns

counts.merge(kin_info)

	kinase_uniprot	substrate_uniprot_site	Entry Name	Protein names	Gene Names	uniprot_keyword_kinase	on_tree	Organism	Keywords	Sequence
0	O00141	387	SGK1_HUMAN	Serine/threonine-protein kinase Sgk1 (EC 2.7.1...	SGK1 SGK	1	1	Homo sapiens (Human)	3D-structure;Alternative promoter usage;Altern...	MTVKTEAAKGTLTYSRMRGMVAILIAFMKQRRMGLNDFIQKIANNS...
1	O00238	356	BMR1B_HUMAN	Bone morphogenetic protein receptor type-1B (B...	BMPR1B	1	1	Homo sapiens (Human)	3D-structure;Alternative splicing;ATP-binding;...	MLLRSAGKLNVGTKKEDGESTAPTPRPKVLRCKCHHHCPEDSVNNI...
2	O00311	50	CDC7_HUMAN	Cell division cycle 7-related protein kinase (...	CDC7 CDC7L1	1	1	Homo sapiens (Human)	3D-structure;Alternative splicing;ATP-binding;...	MEASLGIQMDEPMAFSPQRDRFQAEGSLKKNEQNFKLAGVKKDIEK...
3	O00329	6	PK3CD_HUMAN	Phosphatidylinositol 4,5-bisphosphate 3-kinase...	PIK3CD	1	0	Homo sapiens (Human)	3D-structure;Adaptive immunity;Alternative spl...	MPPGVDCPMEFWTKEENQSVVVDFLLPTGVYLNFPVSRNANLSTIK...
4	O00418	202	EF2K_HUMAN	Eukaryotic elongation factor 2 kinase (eEF-2 k...	EEF2K	1	1	Homo sapiens (Human)	3D-structure;Acetylation;ATP-binding;Calcium;C...	MADEDLIFRLEGVDGGQSPRAGHDGDSDGDSDDEEGYFICPITDDP...
...	...	...	...	...	...	...	...	...	...	...
447	Q9Y572	10	RIPK3_HUMAN	Receptor-interacting serine/threonine-protein ...	RIPK3 RIP3	1	1	Homo sapiens (Human)	3D-structure;Alternative splicing;Apoptosis;AT...	MSCVKLWPSGAPAPLVSIEELENQELVGKGGFGTVFRAQHRKWGYD...
448	Q9Y5S2	109	MRCKB_HUMAN	Serine/threonine-protein kinase MRCK beta (EC ...	CDC42BPB KIAA1124	1	1	Homo sapiens (Human)	3D-structure;ATP-binding;Cell junction;Cell me...	MSAKVRLKKLEQLLLDGPWRNESALSVETLLDVLVCLYTECSHSAL...
449	Q9Y6E0	846	STK24_HUMAN	Serine/threonine-protein kinase 24 (EC 2.7.11....	STK24 MST3 STK3	1	1	Homo sapiens (Human)	3D-structure;Acetylation;Alternative splicing;...	MDSRAQLWGLALNKRRATLPHPGGSTNLKADPEELFTKLEKIGKGS...
450	Q9Y6M4	224	KC1G3_HUMAN	Casein kinase I isoform gamma-3 (CKI-gamma 3) ...	CSNK1G3	1	1	Homo sapiens (Human)	3D-structure;Acetylation;Alternative splicing;...	MENKKKDKDKSDDRMARPSGRSGHNTRGTGSSSSGVLMVGPNFRVG...
451	Q9Y6R4	15	M3K4_HUMAN	Mitogen-activated protein kinase kinase kinase...	MAP3K4 KIAA0213 MAPKKK4 MEKK4 MTK1	1	1	Homo sapiens (Human)	Alternative splicing;ATP-binding;Cytoplasm;Kin...	MREAAAALVPPPAFAVTPAAAMEEPPPPPPPPPPPPEPETESEPEC...

452 rows × 10 columns

Source confidence

df_group.source.str.split('|').str.len().value_counts()

source
1    174536
2      6123
3      3352
4      1312
5       923
6       770
7        50
Name: count, dtype: int64

Self-phosphorylate

df_group[df_group['kinase_uniprot']==df_group['substrate_uniprot']].substrate_genes.value_counts().head(20)

substrate_genes
LRRK2 PARK8                  57
TRPM7 CHAK1 LTRPC7           46
MARK2 EMK1                   35
BMX                          34
TTK MPS1 MPS1L1              34
NEK4 STK2                    33
DCLK2 DCAMKL2 DCDC3B DCK2    32
ERBB4 HER4                   32
ABL1 ABL JTK7                32
MARK1 KIAA1477 MARK          31
HIPK2                        31
RPS6KA2 MAPKAPK1C RSK3       29
BTK AGMX1 ATK BPK            29
STK4 KRS2 MST1               28
BUB1B BUBR1 MAD3L SSK1       28
PDGFRA PDGFR2 RHEPDGFRA      28
ITK EMT LYK                  28
RPS6KA6 RSK4                 28
PRKD2 PKD2 HSPC187           28
ABL2 ABLL ARG                27
Name: count, dtype: int64

df_group.source

source
Sugiyama                     167633
EPSD|PSP                       3346
PSP                            2932
SIGNOR                         1984
SIGNOR|EPSD|PSP                1878
                              ...  
GPS6|SIGNOR|ELM|Sugiyama          1
ELM|iPTMNet|Sugiyama              1
ELM|iPTMNet|EPSD|Sugiyama         1
iPTMNet|PSP|Sugiyama              1
SIGNOR|ELM|EPSD|Sugiyama          1
Name: count, Length: 111, dtype: int64

Nearby co-occured kinase