Preprocess Large-scale dataset

import pandas as pd

df = pd.read_csv('raw/Large_scale_S2.csv')

# Add substrate and phos position info
df['Substrate'] = df.Substrate_uniprot.str.split('_').str[0]
df['substrate_position'] = df['Substrate'] + "_"+ df['Position']

EDA - Pivot table

index is unique phosphosites and column is kinase

p = df.pivot_table(index='substrate_position',
                   columns='Kinase',
                   values='Substrate', # any column
                   aggfunc='count',
                   fill_value=0)

p

Kinase	ABL1	ABL1[E255K]	ABL1[T315I]	ABL2	ACK	ACTR2	ACTR2B	AKT1	AKT2	AKT3	...	ZAK	ZAP70	p38a	p38b	p38d	p38g	p70S6K	p70S6Kb	skMLCK	smMLCK
substrate_position
1433B_S132	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1433B_S145	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1433B_S147	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1433B_S158	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1433B_S212	1	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
ZW10_S611	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
ZYX_S259	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	1	0	0	0	0	0
ZYX_S267	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	1	1	0	0	0	0
ZYX_S505	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
ZYX_T270	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	1	1	0	0	0	0

21449 rows × 385 columns

# Check mutation
for kinase_with_mutation in ['ABL','ALK','BRAF','EGFR','FGFR3',
                             'KIT','LRRK2','MET','PDGFRa','RET']:
    print(p[p.columns[p.columns.str.contains(kinase_with_mutation)]].sum().reset_index(name='#substrates'))
    print('---------------')

        Kinase  #substrates
0         ABL1         1577
1  ABL1[E255K]         1274
2  ABL1[T315I]         1322
3         ABL2         1004
---------------
        Kinase  #substrates
0          ALK         1876
1         ALK1           29
2         ALK2          331
3         ALK4          463
4  ALK[F1174L]         1111
5  ALK[R1275Q]         1044
---------------
        Kinase  #substrates
0         BRAF           22
1  BRAF[V600E]          195
---------------
              Kinase  #substrates
0               EGFR          796
1        EGFR[L858R]          648
2        EGFR[L861Q]          981
3  EGFR[T790M/L858R]         1112
4        EGFR[T790M]         1086
---------------
         Kinase  #substrates
0         FGFR3          775
1  FGFR3[K650E]          827
2  FGFR3[K650M]         1220
---------------
       Kinase  #substrates
0         KIT          786
1  KIT[T670I]          674
2  KIT[V560G]          172
---------------
          Kinase  #substrates
0          LRRK2           59
1  LRRK2[G2019S]           11
---------------
        Kinase  #substrates
0          MET         1457
1  MET[Y1235D]          511
---------------
          Kinase  #substrates
0         PDGFRa         1194
1  PDGFRa[T674I]          249
2  PDGFRa[V561D]         1175
---------------
       Kinase  #substrates
0         RET         1735
1  RET[G691S]         1165
2  RET[M918T]         1221
3  RET[S891A]          876
4  RET[Y791F]          860
---------------

# Number of substrates of kinase
p.sum().sort_values(ascending=False)

Kinase
EPHA3             2102
FES               2008
TRKC              1928
SRC               1899
EPHA8             1878
                  ... 
PKN3                 6
SPHK1                4
PIK3CD/PIK3R1        3
PFTAIRE1/cycD3       2
PRPK                 2
Length: 385, dtype: int64

# Most phosphorylated substrates
p.sum(1).sort_values(ascending=False)

substrate_position
HSPB1_S82     307
ALDOA_S39     191
HS90B_S462    189
RS29_Y7       185
RS16_S9       180
             ... 
ZN768_Y359      1
ZN687_S316      1
ZN706_Y39       1
1433F_T210      1
1433F_T231      1
Length: 21449, dtype: int64

index is unique substrates and column is kinase

p2 = df.pivot_table(index='Substrate',
                   columns='Kinase',
                   values='Type', # any column
                   aggfunc='count',
                   fill_value=0)

p2

Kinase	ABL1	ABL1[E255K]	ABL1[T315I]	ABL2	ACK	ACTR2	ACTR2B	AKT1	AKT2	AKT3	...	ZAK	ZAP70	p38a	p38b	p38d	p38g	p70S6K	p70S6Kb	skMLCK	smMLCK
Substrate
1433B	4	4	3	4	2	0	0	1	0	0	...	0	3	1	1	1	1	0	0	0	0
1433E	3	4	3	4	2	1	0	1	0	0	...	1	4	1	1	2	2	1	0	0	0
1433F	4	3	4	4	1	0	0	1	0	0	...	0	3	1	1	1	2	1	0	0	0
1433G	4	4	4	3	1	0	0	1	0	0	...	1	2	2	2	2	2	1	0	0	0
1433S	5	4	3	3	2	0	0	1	0	0	...	0	3	2	1	2	2	0	0	0	0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
ZO1	2	1	0	1	0	0	0	0	0	0	...	0	0	1	0	0	1	1	0	0	0
ZO2	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
ZRAB2	0	0	2	0	0	0	0	0	0	0	...	2	4	0	0	0	0	1	0	0	0
ZW10	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
ZYX	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	3	2	0	0	0	0

4128 rows × 385 columns

Find substrates that share same kinase pattern

p2[p2.duplicated(keep=False)].sort_values(list(p2.columns))

Kinase	ABL1	ABL1[E255K]	ABL1[T315I]	ABL2	ACK	ACTR2	ACTR2B	AKT1	AKT2	AKT3	...	ZAK	ZAP70	p38a	p38b	p38d	p38g	p70S6K	p70S6Kb	skMLCK	smMLCK
Substrate
ATS1	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	1	0	0	0
CARM1	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	1	0	0	0
CLN3	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	1	0	0	0
DPOE1	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	1	0	0	0
EHMT2	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	1	0	0	0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
PP2AB	4	2	4	1	0	0	0	0	0	0	...	0	0	1	0	0	1	0	0	0	0
ACTB	6	7	5	3	4	2	0	2	2	2	...	11	4	6	4	7	5	3	2	4	0
ACTG	6	7	5	3	4	2	0	2	2	2	...	11	4	6	4	7	5	3	2	4	0
TBA1A	7	4	4	6	1	1	0	0	0	0	...	2	3	5	4	3	4	0	1	0	0
TBA1B	7	4	4	6	1	1	0	0	0	0	...	2	3	5	4	3	4	0	1	0	0

869 rows × 385 columns

ID mapping

Query the uniprot id of substrates in the uniprot web

seq=pd.read_excel('raw/id_mapping.xlsx')

seq = seq.rename(columns={'From':'uniprot','Sequence':'sequence'})

df = pd.read_csv('raw/Large_scale_S2.csv').rename(columns={'Substrate_uniprot':'uniprot'})

df

	Type	Kinase	uniprot	Position	SIDIC	PTMscore\n(P > 075)
0	TK	ABL1	1433B_HUMAN	S212	1	1
1	TK	ABL1	1433B_HUMAN	Y151	1	1
2	TK	ABL1	1433B_HUMAN	Y21	1	1
3	TK	ABL1	1433B_HUMAN	Y50	1	1
4	TK	ABL1	1433E_HUMAN	Y152	1	1
...	...	...	...	...	...	...
198531	LK	SPHK2	TICN3_HUMAN	T118	1	1
198532	LK	SPHK2	TPM4_HUMAN	T241	1	1
198533	LK	SPHK2	ULK3_HUMAN	S305	1	1
198534	LK	SPHK2	ZRAB2_HUMAN	S165	1	1
198535	LK	SPHK2	ZRAB2_HUMAN	S181	1	1

198536 rows × 6 columns

df = df.merge(seq)

# Extract amino acid and position from "Position"
df['phospho_receptor'] = df['Position'].str[0]

df['phospho_position'] = df['Position'].str[1:].astype(int)

df.phospho_receptor.value_counts()

phospho_receptor
Y    106611
S     63242
T     28683
Name: count, dtype: int64

def validate_position(row):
    # Extract amino acid and position from the new columns
    amino_acid = row['phospho_receptor']
    position = int(row['phospho_position'])
    
    try:
        # Check if the amino acid at the given position matches the specified amino acid
        if row['sequence'][position-1] == amino_acid:
            return 1
        else:
            return 0
    except IndexError:  # Handle the case when position-1 exceeds the length of sequence
        return 0

df['Validated'] = df.apply(validate_position, axis=1)

df.Validated.value_counts()

Validated
1    196126
0      2410
Name: count, dtype: int64

df.query('Validated==0').uniprot.value_counts()

uniprot
EFTU_HUMAN     339
DDX17_HUMAN    217
ARI1B_HUMAN    114
ANM1_HUMAN      95
MIER1_HUMAN     87
              ... 
PAR14_HUMAN      1
ZNF30_HUMAN      1
RGS10_HUMAN      1
RAB4A_HUMAN      1
WDR81_HUMAN      1
Name: count, Length: 97, dtype: int64

Find substrate that does not have ideal amino acid at the position

invalid = df.query('Validated==0').uniprot.value_counts().reset_index(name='count')\
.rename(columns={'index':'uniprot'})\
.merge(seq)

invalid

	uniprot	count	sequence
0	EFTU_HUMAN	339	MTTMAAATLLRATPHFSGLAAGRTFLLQGLLRLLKAPALPLLCRGL...
1	DDX17_HUMAN	217	MPTGFVAPILCVLLPSPTREAATVASATGDSASERESAAPAAAPTA...
2	ARI1B_HUMAN	114	MAARAAAAAAAAAARARARAGSGERRAPPGPRPAPGARDLEAGARG...
3	ANM1_HUMAN	95	MAAAEAANCIMENFVATLANGMSLQPPLEEVSCGQAESSEKPNAED...
4	MIER1_HUMAN	87	MAEPSVESSSPGGSATSDDHEFDPSADMLVHDFDDERTLEEEEMME...
...	...	...	...
92	PAR14_HUMAN	1	MAVPGSFPLLVEGSWGPDPPKNLNTKLQMYFQSPKRSGGGECEVRQ...
93	ZNF30_HUMAN	1	MAHKYVGLQYHGSVTFEDVAIAFSQQEWESLDSSQRGLYRDVMLEN...
94	RGS10_HUMAN	1	MFNRAVSRLSRKRPPSDIHDSDGSSSSSHQSLKSTAKWAASLENLL...
95	RAB4A_HUMAN	1	MSQTAMSETYDFLFKFLVIGNAGTGKSCLLHQFIEKKFKDDSNHTI...
96	WDR81_HUMAN	1	MAQGSGGREGALRTPAGGWHSPPSPDMQELLRSVERDLSIDPRQLA...

97 rows × 3 columns

Replace the invalidated sequence with updated sequence

# updated with valid sequences from old uniprot database
update1 = pd.read_excel('raw/seq.xlsx')

# full sequence
seq = pd.read_csv('raw/large_scale_sequence.csv').drop(columns=['uniprot_updated'])

# get a new column "sequence_new" that shows the updated sequence
result = seq.merge(update1,on='uniprot',how='left',suffixes=('', '_new'))

# Where a new sequence is available, replace the old sequence
result['sequence'] = result['sequence_new'].combine_first(result['sequence'])

# Drop the auxiliary columns
result = result.drop(columns=['sequence_new'])

seq = result.copy()

seq

	uniprot	sequence
0	1433B_HUMAN	MTMDKSELVQKAKLAEQAERYDDMAAAMKAVTEQGHELSNEERNLL...
1	1433E_HUMAN	MDDREDLVYQAKLAEQAERYDEMVESMKKVAGMDVELTVEERNLLS...
2	1433F_HUMAN	MGDREQLLQRARLAEQAERYDDMASAMKAVTELNEPLSNEDRNLLS...
3	1433G_HUMAN	MVDREQLVQKARLAEQAERYDDMAAAMKNVTELNEPLSNEERNLLS...
4	1433S_HUMAN	MERASLIQKAKLAEQAERYEDMAAFMKGAVEKGEELSCEERNLLSV...
...	...	...
4123	1B42_HUMAN	MLVMAPRTVLLLLSAALALTETWAGSHSMRYFYTSVSRPGRGEPRF...
4124	FA21D_HUMAN	MRGKRRPQTRAARRLAAQESSEAEDMSVPRGPIAQWADGAISPNGH...
4125	1C06_HUMAN	MRVMAPRTLILLLSGALALTETWACSHSMRYFDTAVSRPGRGEPRF...
4126	MPP6_HUMAN	MQQVLENLTELPSSTGAEEIDLIFLKGIMENPIVKSLAKAHERLED...
4127	1C16_HUMAN	MRVMAPRALLLLLSGGLALTETWACSHSMRYFDTAVSRPGRGEPRF...

4128 rows × 2 columns

Validate again

import pandas as pd

df = pd.read_csv('raw/Large_scale_S2.csv').rename(columns={'Substrate_uniprot':'uniprot'})

df = df.merge(result)

# Extract amino acid and position from "Position"
df['phospho_receptor'] = df['Position'].str[0]
df['phospho_position'] = df['Position'].str[1:].astype(int)

def validate_position(row):
    # Extract amino acid and position from the new columns
    amino_acid = row['phospho_receptor']
    position = int(row['phospho_position'])
    
    try:
        # Check if the amino acid at the given position matches the specified amino acid
        if row['sequence'][position-1] == amino_acid:
            return 1
        else:
            return 0
    except IndexError:  # Handle the case when position-1 exceeds the length of sequence
        return 0

df['Validated'] = df.apply(validate_position, axis=1)

# invalidated
df.query('Validated==0').shape

(215, 10)

Much less than before (2410) !

invalid = df.query('Validated==0')

# remove uniprot that has invalid phosphoreceptor, and drop sequence column
df = df[~df.uniprot.isin(invalid.uniprot)].drop(columns=['sequence'])

# as we modify the sequence, we need to replace with new sequence
df = df.merge(result)

df

	Type	Kinase	uniprot	Position	SIDIC	PTMscore\n(P > 075)	phospho_receptor	phospho_position	Validated	sequence
0	TK	ABL1	1433B_HUMAN	S212	1	1	S	212	1	MTMDKSELVQKAKLAEQAERYDDMAAAMKAVTEQGHELSNEERNLL...
1	TK	ABL1	1433B_HUMAN	Y151	1	1	Y	151	1	MTMDKSELVQKAKLAEQAERYDDMAAAMKAVTEQGHELSNEERNLL...
2	TK	ABL1	1433B_HUMAN	Y21	1	1	Y	21	1	MTMDKSELVQKAKLAEQAERYDDMAAAMKAVTEQGHELSNEERNLL...
3	TK	ABL1	1433B_HUMAN	Y50	1	1	Y	50	1	MTMDKSELVQKAKLAEQAERYDDMAAAMKAVTEQGHELSNEERNLL...
4	TK	ABL2	1433B_HUMAN	Y106	1	1	Y	106	1	MTMDKSELVQKAKLAEQAERYDDMAAAMKAVTEQGHELSNEERNLL...
...	...	...	...	...	...	...	...	...	...	...
198243	LK	SPHK2	SPHK2_HUMAN	T389	1	1	T	389	1	MNGHLEAEEQQDQRPDQELTGSWGHGPRSTLVRAKAMAPPPPPLAA...
198244	LK	SPHK2	SPHK2_HUMAN	T402	1	1	T	402	1	MNGHLEAEEQQDQRPDQELTGSWGHGPRSTLVRAKAMAPPPPPLAA...
198245	LK	SPHK2	SPHK2_HUMAN	T404	1	1	T	404	1	MNGHLEAEEQQDQRPDQELTGSWGHGPRSTLVRAKAMAPPPPPLAA...
198246	LK	SPHK2	SPHK2_HUMAN	T503	1	1	T	503	1	MNGHLEAEEQQDQRPDQELTGSWGHGPRSTLVRAKAMAPPPPPLAA...
198247	LK	SPHK2	SPHK2_HUMAN	T614	1	1	T	614	1	MNGHLEAEEQQDQRPDQELTGSWGHGPRSTLVRAKAMAPPPPPLAA...

198248 rows × 10 columns

Convert substrate sequence to lower case

# for each substrate, find their phosphorylation point
modify = df.groupby('uniprot').agg({'Position':lambda r: r.unique()}).reset_index()

modify = modify.merge(result)

modify

	uniprot	Position	sequence
0	1433B_HUMAN	[S212, Y151, Y21, Y50, Y106, Y213, S47, S39, T...	MTMDKSELVQKAKLAEQAERYDDMAAAMKAVTEQGHELSNEERNLL...
1	1433E_HUMAN	[Y152, Y20, Y49, Y214, S148, S46, Y131, Y121, ...	MDDREDLVYQAKLAEQAERYDEMVESMKKVAGMDVELTVEERNLLS...
2	1433F_HUMAN	[S215, Y154, Y20, Y49, S46, S150, Y216, S145, ...	MGDREQLLQRARLAEQAERYDDMASAMKAVTELNEPLSNEDRNLLS...
3	1433G_HUMAN	[S215, Y20, Y216, Y49, S46, S38, T31, S64, Y12...	MVDREQLVQKARLAEQAERYDDMAAAMKNVTELNEPLSNEERNLLS...
4	1433S_HUMAN	[S149, Y151, Y19, Y213, Y48, S45, S209, S37, T...	MERASLIQKAKLAEQAERYEDMAAFMKGAVEKGEELSCEERNLLSV...
...	...	...	...
4094	ZO1_HUMAN	[Y1165, Y669, S1064, Y1354, Y1355, Y1066, S106...	MSARAAAAKSTAMEETAIWEQHTVTLHRAPGFGFGIAISGGRDNPH...
4095	ZO2_HUMAN	[Y506, S430, Y426, Y1166]	MPVRGDRGFPPRRELSGWLRAPGMEELIWEQYTVTLQKDSKRGFGI...
4096	ZRAB2_HUMAN	[S188, T100, Y167, S120, Y114, Y124, S181, S15...	MSTKNFRVSDGDWICPDKKCGNVNFARRTSCNRCGREKTTEAKMMK...
4097	ZW10_HUMAN	[S605, S611]	MASFVTEVLAHSGRLEKEDLGTRISRLTRRVEEIKGEVCNMISKKY...
4098	ZYX_HUMAN	[S505, S259, S267, T270]	MAAPRPSPAISVSVSAPAFYAPQKKFGPVVAPKPKVNPFRPGDSEP...

4099 rows × 3 columns

# Modify sequence based on position
def modify_sequence(row):
    seq = list(row['sequence'])
    for pos in row['Position']:
        # Extract character and position
        char = pos[0]
        position = int(pos[1:]) - 1  # Subtracting 1 because Python uses 0-based indexing
        
        # Check if the character at the position matches
        # if seq[position] == char:
        seq[position] = char.lower()
    return ''.join(seq)

modify['sequence2'] = modify.apply(modify_sequence, axis=1)

seq2 = modify[['uniprot','sequence2']]

# merge with transformed sequence
df = df.drop(columns=['sequence']).merge(seq2)

df.Kinase.value_counts()

Kinase
EPHA3             2101
FES               2007
TRKC              1926
SRC               1896
EPHA8             1877
                  ... 
PKN3                 6
SPHK1                4
PIK3CD/PIK3R1        3
PFTAIRE1/cycD3       2
PRPK                 2
Name: count, Length: 385, dtype: int64

Extract phosphosite sequence

data = []
for i, r in df.iterrows():
    position = r.phospho_position - 1
    start = position - 7
    end = position + 8

    # Extract the subsequence
    subseq = r.sequence2[max(0, start):min(len(r.sequence2), end)]

    # Pad the subsequence if needed
    if start < 0:
        subseq = "_" * abs(start) + subseq
    if end > len(r.sequence2):
        subseq = subseq + "_" * (end - len(r.sequence2))
    
    data.append(subseq)
    # break

df['substrate'] = data

# check if the middle position belongs to s,t,y
df.substrate.str[7].value_counts()

substrate
y    106558
s     63030
t     28660
Name: count, dtype: int64

# check if the middle position belongs to s,t,y
df.substrate.apply(len).value_counts()

substrate
15    198248
Name: count, dtype: int64

df.substrate.value_counts()[:10]

substrate
IEQEGPEyWDRNTQI    1085
RKEsysVyVyKVLKQ     765
SSSLEKsyELPDGQV     765
VGMGQKDsyVGDEAQ     750
EsysVyVyKVLKQVH     729
TLNNKFAsFIDKVRF     708
MWISKQEyDEsGPsI     693
LPDGQVItIGNERFR     610
EyDEsGPsIVHRKCF     594
QGNRttPsyVAFtDt     556
Name: count, dtype: int64

df['KS_pairs'] = df.Kinase+"_"+df.uniprot.str.split('_').str[0]+"_"+df.Position

df['S_position'] = df.uniprot.str.split('_').str[0]+"_"+df.Position

df['Uniprot'] = df.uniprot.str.split('_').str[0]

Check the most frequence substrate sequence

df.query('substrate == "RKEsysVyVyKVLKQ"').S_position.value_counts()

S_position
H2B1C_Y41    85
H2B1D_Y41    85
H2B1H_Y41    85
H2B1K_Y41    85
H2B1L_Y41    85
H2B1M_Y41    85
H2B1N_Y41    85
H2B2F_Y41    85
H2BFS_Y41    85
Name: count, dtype: int64

Check if the sequences of these substrates are identical

name = df.query('substrate == "RKEsysVyVyKVLKQ"').S_position.value_counts().index

df[df.S_position.isin(name)].sequence2.unique()

array(['MPEPAKSAPAPKKGSKKAVTKAQKKDGKKRKRSRKEsysVyVyKVLKQVHPDTGISSKAMGIMNSFVNDIFERIAGEASRLAHYNKRSTITSREIQTAVRLLLPGELAKHAVSEGTKAVTKYTSSK',
       'MPEPTKSAPAPKKGSKKAVTKAQKKDGKKRKRSRKEsysVyVyKVLKQVHPDTGISSKAMGIMNSFVNDIFERIAGEASRLAHYNKRSTITSREIQTAVRLLLPGELAKHAVSEGTKAVTKYTSSK',
       'MPDPAKSAPAPKKGSKKAVTKAQKKDGKKRKRSRKEsysVyVyKVLKQVHPDTGISSKAMGIMNSFVNDIFERIAGEASRLAHYNKRSTITSREIQTAVRLLLPGELAKHAVSEGTKAVTKYTSSK',
       'MPEPAKSAPAPKKGSKKAVTKAQKKDGKKRKRSRKEsysVyVyKVLKQVHPDTGISSKAMGIMNSFVNDIFERIAGEASRLAHYNKRSTITSREIQTAVRLLLPGELAKHAVSEGTKAVTKYTSAK',
       'MPELAKSAPAPKKGSKKAVTKAQKKDGKKRKRSRKEsysVyVyKVLKQVHPDTGISSKAMGIMNSFVNDIFERIASEASRLAHYNKRSTITSREIQTAVRLLLPGELAKHAVSEGTKAVTKYTSSK',
       'MPEPVKSAPVPKKGSKKAINKAQKKDGKKRKRSRKEsysVyVyKVLKQVHPDTGISSKAMGIMNSFVNDIFERIAGEASRLAHYNKRSTITSREIQTAVRLLLPGELAKHAVSEGTKAVTKYTSSK',
       'MPEPSKSAPAPKKGSKKAVTKAQKKDGKKRKRSRKEsysVyVyKVLKQVHPDTGISSKAMGIMNSFVNDIFERIAGEASRLAHYNKRSTITSREIQTAVRLLLPGELAKHAVSEGTKAVTKYTSSK',
       'MPDPAKSAPAPKKGSKKAVTKVQKKDGKKRKRSRKEsysVyVyKVLKQVHPDTGISSKAMGIMNSFVNDIFERIAGEASRLAHYNKRSTITSREIQTAVRLLLPGELAKHAVSEGTKAVTKYTSSK',
       'MPEPAKSAPAPKKGSKKAVTKAQKKDGRKRKRSRKEsysVyVyKVLKQVHPDTGISSKAMGIMNSFVNDIFERIAGEASRLPHYNKRSTITSREIQTAVRLLLPGELAKHAVSEGTKAVTKYTSAK'],
      dtype=object)

Save

df.shape

(198248, 14)

# df.to_excel('raw/large_scale_final2.xlsx',index=False)