Note
Go to the end to download the full example code
Plot epitope mapping data onto protein sequence alignments¶
Peptide arrays can be used as a high-throughput platform for screening biological interactions. Typical screenings involve the immobilization of diverse peptides on a solid surface to study their interactions with various target molecules. Specifically, arrays of peptides with overlapping sequences can be used to identify the epitope of antibodies on a protein antigen at amino acid level.
General scannings for molecular recognition using peptide arrays
are particlularly useful for epitope identification on monoclonal
antibodies. This example visualizes the data from two epitope mapping
studies, using a color coded sequence alignment representation
of the antigens screened. The scannings interrogated a monoclonal
antibody (MAb) against two arrays of overlaping peptides 1.
The files containing peptide array data can be downloaded
here
and
here.
The antigens screened span the extracellular domain of VAR2CSA, a
virulence factor of Plasmodiun falciparum for the strains FCR3
(residues 1-2659) and NF54 (residues 1-2652). The sequence of
the two domains can be downloaded
here.
First, we generate a sequence aligment of the two VAR2CSA strains:
# Code source: Daniel Ferrer-Vinals
# License: BSD 3 clause
import matplotlib as mpl
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import biotite.sequence as seq
import biotite.sequence.align as align
import biotite.sequence.graphics as graphics
import biotite.sequence.io.fasta as fasta
# Path to the data files
array_seq_path = "../../download/Array_Seq.txt"
fasta_file = fasta.FastaFile.read(array_seq_path)
# Parse protein sequences of FCR3 and NF54
for name, sequence in fasta_file.items():
if "AAQ73926" in name:
fcr3_seq = seq.ProteinSequence(sequence)
elif "EWC87419" in name:
nf54_seq = seq.ProteinSequence(sequence)
# Get BLOSUM62 matrix
matrix = align.SubstitutionMatrix.std_protein_matrix()
# Perform pairwise sequence alignment
alignments = align.align_optimal(fcr3_seq, nf54_seq, matrix,
gap_penalty = (-10, -1),
terminal_penalty = False)
alignment = alignments[0]
print(alignment)
GSGSGSGMDSTSTIANKIEEYLGAKSDDSKIDELLKADPSEVEYYRSGGDGDYLKNNICKITVNHSDSGK
GSGSGSGMDKSS-IANKIEAYLGAKSDDSKIDQSLKADPSEVQYYGSGGDGYYLRKNICKITVNHSDSGT
YDPCEKKLPPYDDNDQWKCQQNSSDGSGKPENICVPPRRERLCTYNLENLKFDKIRDNNAFLADVLLTAR
NDPCDRIPPPYGDNDQWKCAIILSKVSEKPENVFVPPRRQRMCINNLEKLNVDKIRDKHAFLADVLLTAR
NEGEKIVQNHPDTNSSNVCNALERSFADLADIIRGTDQWKGTNSNLEKNLKQMFAKIRENDKVLQDKYPK
NEGERIVQNHPDTNSSNVCNALERSFADIADIIRGTDLWKGTNSNLEQNLKQMFAKIRENDKVLQDKYPK
DQKYTKLREAWWNANRQKVWEVITCGARSNDLLIKRGWRTSGKSDRKKNFELCRKCGHYEKEVPTKLDYV
DQNYRKLREDWWNANRQKVWEVITCGARSNDLLIKRGWRTSGKSNGDNKLELCRKCGHYEEKVPTKLDYV
PQFLRWLTEWIEDFYREKQNLIDDMERHREECTREDHKSKEGTSYCSTCKDKCKKYCECVKKWKTEWENQ
PQFLRWLTEWIEDFYREKQNLIDDMERHREECTSEDHKSKEGTSYCSTCKDKCKKYCECVKKWKSEWENQ
ENKYKDLYEQNKNKTSQKNTSRYDDYVKDFFEKLEANYSSLENYIKGDPYFAEYATKLSFILNPSDANNP
KNKYTELYQQNKNETSQKNTSRYDDYVKDFFKKLEANYSSLENYIKGDPYFAEYATKLSFILNSSDANNP
SGETANHNDEACNCNESGISSVGQAQTSGPSSNKTCITHSSIKTNKKKECKDVKLGVRENDKDLKICVIE
SEKIQKNNDEVCNCNESGIASVEQEQISDPSSNKTCITHSSIKANKKKVCKHVKLGVRENDKDLRVCVIE
DTSLSGVDNCCCQDLLGILQENCSDNKRGSSSNDSCDNKNQDECQKKLEKVFASLTNGYKCDKCKSGTSR
HTSLSGVENCCCQDFLRILQENCSDNKSGSSSNGSCNNKNQEACEKNLEKVLASLTNCYKCDKCKSEQSK
-SKKKWIWKKSSGNEEGLQEEYANTIGLPPRTQSLYLGN-LPKLENVCEDVKDINFDTKEKFLAGCLIVS
KNNKNWIWKKSSGKEGGLQKEYANTIGLPPRTQSLCLVVCLDEKGKKTQELKNIR--TNSELLKEWIIAA
FHEGKNLKKRYPQNKNSGNKENLCKALEYSFADYGDLIKGTSIWDNEYTKDLELNLQNNFGKLFGKYIKK
FHEGKNLKPSH-EKKNDDNGKKLCKALEYSFADYGDLIKGTSIWDNEYTKDLELNLQKIFGKLFRKYIKK
NNTAEQDTSYSSLDELRESWWNTNKKYIWTAMKHGAEMNITTCNADGSVTGSGSSCDDIPTIDLIPQYLR
NNTAEQDTSYSSLDELRESWWNTNKKYIWLAMKHGAGMNSTTCCGDGSVTGSGSSCDDIPTIDLIPQYLR
FLQEWVENFCEQRQAKVKDVITNCKSCKESGNKCKTECKTKCKDECEKYKKFIEACGTAGGGIGTAGSPW
FLQEWVEHFCKQRQEKVKPVIENCKSCKESGGTCNGECKTECKNKCEVYKKFIEDCK---GGDGTAGSSW
SKRWDQIYKRYSKHIEDAKRNRKAGTKNCGTSSTTNAAASTDENKCVQSDIDSFFKHLIDIGLTTPSSYL
VKRWDQIYKRYSKYIEDAKRNRKAGTKNCGPSSTTNAA----ENKCVQSDIDSFFKHLIDIGLTTPSSYL
SNVLDDNICGADKAPWTTYTTYTTTEKCNKERDKSKSQSSDTLVVVNVPSPLGNTPYRYKYACQCKIPTN
SIVLDDNICGADKAPWTTYTTYTTTEKCNKETDKSKLQQCNTAVVVNVPSPLGNTPHGYKYACQCKIPTN
EETCDDRKEYMNQWSCGSARTMKRGYKNDNYELCKYNGVDVKPTTVRSNSSKLDGNDVTFFNLFEQWNKE
EETCDDRKEYMNQWSCGSARTMKRGYKNDNYELCKYNGVDVKPTTVRSNSSKLDDKDVTFFNLFEQWNKE
IQYQIEQYMTNANISCIDEKEVLDSVSDEGT-PKVRGGYEDGRNNNTDQGTNCKEKCKCYKLWIEKINDQ
IQYQIEQYMTNTKISCNNEKNVLSRVSDEAAQPKFSDNERD-RNSITHEDKNCKEKCKCYSLWIEKINDQ
WGKQKDNYNKFRSKQIYDANKGSQNKKVVSLSNFLFFSCWEEYIQKYFNGDWSKIKNIGSDTFEFLIKKC
WDKQKDNYNKFQRKQIYDANKGSQNKKVVSLSNFLFFSCWEEYIQKYFNGDWSKIKNIGSDTFEFLIKKC
GNNSAHGEEIFNEKLKNAEKKCKENESTDTNINKSETSCDLNATNYIRGCQSKTYDGKIFPGKGGEKQWI
GNDSGDGETIFSEKLNNAEKKCKENESTNNKMKSSETSCDCSEPIYIRGCQPKIYDGKIFPGKGGEKQWI
CKDTIIHGDTNGACIPPRTQNLCVGELWDKSYGGRSNIKNDTKELLKEKIKNAIHKETELLYEYHDTGTA
CKDTIIHGDTNGACIPPRTQNLCVGELWDKRYGGRSNIKNDTKESLKQKIKNAIQKETELLYEYHDKGTA
IISKNDKKGQKGK----NDPNGLPKGFCHAVQRSFIDYKNMILGTSVNIYEHIGKLQEDIKKIIEKGTPQ
IISRNPMKGQKEKEEKNNDSNGLPKGFCHAVQRSFIDYKNMILGTSVNIYEYIGKLQEDIKKIIEKGTTK
QKDKIGGVGSSTENVNAWWKGIEREMWDAVRCAITKINKKNN-NSIFNGDECGVSPPTGNDEDQSVSWFK
QNGKT--VGSGAENVNAWWKGIEGEMWDAVRCAITKINKKQKKNGTFSIDECGIFPPTGNDEDQSVSWFK
EWGEQFCIERLRYEQNIREACTINGKNEKKCINSKSGQGDKIQGACKRKCEKYKKYISEKKQEWDKQKTK
EWSEQFCIERLQYEKNIRDACTNNG------------QGDKIQGDCKRKCEEYKKYISEKKQEWDKQKTK
YENKYVGKSASDLLKENYPECISANFDFIFNDNIEYKTYYPYGDYSSICSCEQVKYYKYNNAEKKNNKSL
YENKYVGKSASDLLKENYPECISANFDFIFNDNIEYKTYYPYGDYSSICSCEQVKYYEYNNAEKKNNKSL
CYEKDNDMTWSKKYIKKLENGRSLEGVYVPPRRQQLCLYELFPIIIKNEEGMEKAKEELLETLQIVAERE
CHEKGNDRTWSKKYIKKLENGRTLEGVYVPPRRQQLCLYELFPIIIKNKNDITNAKKELLETLQIVAERE
AYYLWKQYNPTGKGIDDANKKACCAIRGSFYDLEDIIKGNDLVHDEYTKYIDSKLNEIFGSSDTNDIDTK
AYYLWKQYHAHNDTTYLAHKKACCAIRGSFYDLEDIIKGNDLVHDEYTKYIDSKLNEIFDSSNKNDIETK
RARTDWWENETIT-------NGTDRKTIRQLVWDAMQSGVRYAVEEK------NENFPLCMGVEHIGIAK
RARTDWWENEAIAVPNITGANKSDPKTIRQLVWDAMQSGVRKAIDEEKEKKKPNENFPPCMGVQHIGIAK
PQFIRWLEEWTNEFCEKYTKYFEDMKSKCDPPKRADTCGDNSNIECKKACANYTNWLNPKRIEWNGMSNY
PQFIRWLEEWTNEFCEKYTKYFEDMKSNCNLRKGADDCDDNSNIECKKACANYTNWLNPKRIEWNGMSNY
YNKIYRKSNKESEGGKDYSMIMAPTVIDYLNKRCHGEINGNYICCSCKNIGAYNTTSGTVNKKLQKKETE
YNKIYRKSNKESEDGKDYSMIMEPTVIDYLNKRCNGEINGNYICCSCKNIGE-NSTSGTVNKKLQKKETQ
CEEEKGPLDLMNEVLNKMDKKYSAHKMKCTEVYLEHVEEQLNEIDNAIKDYKLYPLDRCFDDQTKMKVCD
CEDNKGPLDLMNKVLNKMDPKYSEHKMKCTEVYLEHVEEQLKEIDNAIKDYKLYPLDRCFDDKSKMKVCD
LIADAIGCKDKTKLDELDEWNDMDLRGTYNKHKGVLIPPRRRQLCFSRIVRGPANLRSLNEFKEEILKGA
LIGDAIGCKHKTKLDELDEWNDVDMRDPYNKYKGVLIPPRRRQLCFSRIVRGPANLRNLKEFKEEILKGA
QSEGKFLGNYYKEHKDKEKALEAMKNSFYDYEDIIKGTDMLTNIEFKDIKIKLDRLLEKETNNTKKAEDW
QSEGKFLGNYYNEDKDKEKALEAMKNSFYDYEYIIKGSDMLTNIQFKDIKRKLDRLLEKETNNTEKVDDW
WKTNKKSIWNAMLCGYKKSGNKIIDPSWCTIPTTETPPQFLRWIKEWGTNVCIQKQEHKEYVKSKCSNVT
WETNKKSIWNAMLCGYKKSGNKIIDPSWCTIPTTETPPQFLRWIKEWGTNVCIQKEEHKEYVKSKCSNVT
NLGAQASESNNCTSEIKKYQEWSRKRSIRWETISKRYKKYKRMD----ILKDVKEPDANT-----YLREH
NLGAQESESKNCTSEIKKYQEWSRKRSIQWEAISEGYKKYKGMDEFKNTFKNIKEPDANEPNANEYLKKH
CSKCPCGFNDMEEMNNNEDNEKEAFKQIKEQVKIPAELEDVIYRIKHHEYDKGNDYICNKYKNIHDRMKK
CSKCPCGFNDMQEITKYTNIGNEAFKQIKEQVDIPAELEDVIYRLKHHEYDKGNDYICNKYKNINVNMKK
NNGNFVTDNFVKKSWEISNGVLIPPRRKNLFLYIDPSKICEYKKDPKLFKDFIYWSAFTEVERLKKAYGG
NNDDTWTD-LVKNSSDINKGVLLPPRRKNLFLKIDESDICKYKRDPKLFKDFIYSSAISEVERLKKVYGE
ARAKVVHAMKYSFTDIGSIIKGDDMMEKNSSDKIGKILGDTDGQNEKRKKWWDMNKYHIWESMLCGYREA
AKTKVVHAMKYSFADIGSIIKGDDMMENNSSDKIGKILGDGVGQNEKRKKWWDMNKYHIWESMLCGYKHA
EGDTETNEN--CRFPDIESVPQFLRWFQEWSENFCDRRQKLYDKLNSECISAEC--TNGSVDNSKCTHAC
YGNISENDRKMLDIPNNDDEHQFLRWFQEWTENFCTKRNELYENMVTACNSAKCNTSNGSVDKKECTEAC
VNYKNYILTKKTEYEIQTNKYDNEFKNKNSNDKDAPDYLKEKCNDNKCECLNKHIDDKNKTWKNPYETLE
KNYSNFILIKKKEYQSLNSQYDMNYKETKAEKKESPEYFKDKCN-GECSCLSEYFKDETR-WKNPYETLD
DT-FKSKCDCPKPLPSPIKPDDLPPQADEPFDPTILQTTIPGSGSGSG
DTEVKNNCMCKPP----------PPASNNTSD--ILQKTIPGSGSGSG
Epitope mapping data¶
This study used arrays of overlaping peptides to achive high acurracy in mapping the epitope. Both FCR3 and NF54 arrays, consisted of 20-mer peptides with an overlap of 19 and 18 amino acids respectively. Arbitrary units (AU) of fluorescence intensity quantified the antibody recognition for each peptide. Our goal is to decorate the aligment, with the fluorescence intensity scores of each peptide in the arrays. We used a color code from red to white for high to low intensity, respectively. The background color of the symbols on the aligment corresponds to the score for the 20th amino acid at the end of the peptide.
Lets create a function that maps the peptide score to the 20th residue of the peptide:
def read_scan(filename, pep_len=20, score_res=20):
if not type(pep_len) is int:
raise TypeError("pep_len : only integers are allowed")
elif not type(score_res) is int:
raise TypeError("score_res : only integers are allowed")
elif pep_len < score_res:
raise Exception("score_res can't be higher than pep_len")
elif pep_len != 20 or score_res != 20:
s = (score_res) - pep_len -1
else:
s =-1
df= pd.read_csv(filename)
scor_res = df['Seq'].str[s]
df['s_res'] = scor_res
return df
# Load epitope scan data
fcr3_file_path = "../../download/FCR3_10ug.csv"
nf54_file_path = "../../download/NF54_10ug.csv"
# Define the score residues on the arrays
files = [fcr3_file_path, nf54_file_path]
d = 0
for f in files:
if f == files[0]:
ag1_scan = read_scan(files[d], 20, 20)
elif f == files[1]:
ag2_scan = read_scan(files[d], 20, 20)
d = d + 1
ag1_scan.head(5)
The microarrays contained each peptide printed in duplicated spots. We need to combine the values of those experimental replicates into a unique score for each peptide. Typically, this unique value could come from the geometric mean between replicates that do not deviate wildly. If the average deviation between replicates is high, one can assumme that experimental errors should result in a lower score at a given spot. It is easy to imagine that imperfections on the printing of the spot, will rather decrease and not increase, the antibody recognition, in which case the the peptide signal is better represented by the higher score replicate.
Now lets write a function to combine the scores adding the flexibility to choose cases for those criterias exposed above. We will flag with 0 or 1 every peptide entry on the arrays: 1 if the deviation between replicates is higher than 40%, otherwise 0.
def combine_scores(dataframe, combine='max', flag_noisy=True):
df= dataframe
# mean
df['ave'] = df.iloc[:,[1,2]].mean(axis = 1)
# mean deviation
df['avedev'] = ((df.r1 - df.ave).abs() + (df.r2 - df.ave).abs()) / 2
# percent deviation between replicates
df['dev_ratio'] = df.apply(lambda x:0
if x.avedev==0 else x.avedev/x.ave, axis=1)
# signal value:
if combine == 'max':
df['comb_signal'] = df.apply(lambda x:max(x.r1, x.r2)
if x.dev_ratio >=0.4 else x.ave, axis=1)
elif combine == 'mean':
df['comb_signal'] = df.apply(lambda x:x.ave
if x.dev_ratio <= 0.4 else 0, axis=1)
if flag_noisy:
df['flag'] = df.apply(lambda x:0
if x.dev_ratio <= 0.4 else 1, axis=1)
return df
# Make the corresponding signal equal the replicate with the higest
# score value.
dfa = combine_scores(ag1_scan, combine = 'max', flag_noisy = True)
dfb = combine_scores(ag2_scan, combine = 'max', flag_noisy = True)
dfa.head(5)
Many molecular recognition screening campaings e.g. epitope mapping screenings follow a long-tailed data distribution. To properly represent such distribution one can normalize the date using linear or non-linear transformations on the combined score data.
def data_transform(dataframe, threshold=0):
df = dataframe
# Option to set a "threshold" for the signal scores.
t = threshold
df['cubic'] = df.apply(lambda x: np.cbrt(max(0, x.comb_signal-t)),
axis=1)
df['signal_plot'] = df.apply(lambda x: x.cubic/df['cubic'].max(),
axis=1)
# Normalize, using the power law with cubic exponent. No threshold
data_transform(dfa, threshold=0)
data_transform(dfb, threshold=0)
dfa.head(5)
Convert score residues from the epitope scan to alignment-like gapped sequences¶
So far, we have the peptide score data combined, normalized, and mapped to a residue for each peptide. Next, using the alignment trace as a template, we will match the signal intensities associated to the score residues, to the position of each symbol on the alignment, considering the gaps.
# Get the trace for each sequence on the alignment:
trace_a = align.get_symbols(alignment)[0]
trace_b = align.get_symbols(alignment)[1]
def gapped_seq(dataframe, seq_trace, p_len, overlap_step=1):
"""
Generate a gapped sequence that relates peptide score data signal with a
template alignment trace. The function returns a list of tuples representing
the gapped sequence, where each tuple consists of a residue and its associated
signal value.
Parameters
----------
dataframe : DataFrame
A *Pandas* dataframe containing columns for each peptide score data,
and its designated score residue.
seq_trace : list
The sequence trace obtained from the alignment.
p_len : int
The length of each overlapping peptide.
overlap_step : int, optional
The step size for overlapping peptides.Default is 1.
Note:
-----
The 'gapped' sequence may be shorter than the aligment trace if the alignment results
in gaps at either end. Any remaining elements in the trace with 'None' values are
filled with tuples: ('None', 0).
"""
template = seq_trace
df = dataframe
step = overlap_step
gapped = list(zip(df.s_res , df.signal_plot))
lk1 = df["s_res"].values.tolist()
plen = p_len # peptide length
if step == 1:
x, b = 0, 0
c = 0 # cyclic counter up to the peptide length :20
p = 0 # peptide counter
for b in range(len(lk1)):
for a in template[x:]:
if c < plen-1 :
if a==None:
gapped.insert(x,(template[x],0))
x=x+1
elif a != lk1[b]:
gapped.insert(x,(template[x],0))
x=x+1
c=c+1
elif p==0:
gapped.insert(x,(template[x],0))
x=x+1
c=c+1
else:
x=x+1
c=c+1
break
else:
c = 0 # reset the counter
p=p+1
x=x+1
break
elif step == 2:
x, b = 0, 0
c=0
p=0
for b in range(len(lk1)):
for a in template[x:]:
if c < plen-1 and p==0:
if a==None:
gapped.insert(x,(template[x],0))
x=x+1
else:
gapped.insert(x,(template[x],0))
x=x+1
c=c+1
elif p==0 :
c = 0 # reset the counter
p=p+1
x=x+1
break
if p!=0:
if a==None and c == 0:
gapped.insert(x,(template[x],0))
x=x+1
elif c % 2 == 0:
if a==None:
gapped.insert(x,(template[x],0))
x=x+1
else:
gapped.insert(x,(template[x],0))
x=x+1
c=c+1
elif c % 2 != 0:
if a==None:
gapped.insert(x,(template[x],0))
x=x+1
elif a != lk1[b]:
gapped.insert(x,(template[x],0))
x=x+1
c=c+1
else:
x=x+1
c=c+1
break
# For terminal gaps
if len(gapped) < len(template) and template[len(gapped)+1]== None:
gapped_tail=[]
for n in range(len(template)-len(gapped)):
gapped_tail.append(('None', 0))
gapped = gapped + gapped_tail
return gapped
# Let's use gapped_seq() to build the gapped sequences
# FCR3 array, overlap_step: 1 (pep = 20-mer with 19 overlap)
gapd_s1 = gapped_seq(dfa, trace_a, 20, 1)
# NF54 array, overlap_step: 2 (pep = 20-mer with 18 overlap)
gapd_s2 = gapped_seq(dfb, trace_b, 20, 2)
# Checkpoint. Both gapped sequences must have the same length.
len(gapd_s1) == len(gapd_s2)
True
Create a signal map¶
Now we will generate an object mapping the signal scores from two gapped sequences.
def signal_map(gapped_seq1, gapped_seq2,):
"""
Generate a mapping of signal scores from two gapped sequences.
This function takes two gapped sequences, `gapped_seq1` and
`gapped_seq2`. Each sequence is represented as a list of tuples,
with the first element being an amino acid symbol and the second
element being a signal score. It extracts the signal scores from
each sequence and creates a 2D array with two columns, where the
first column contains signal scores from `gapped_seq1` and the
second column contains signal scores from `gapped_seq2`.
Parameters:
-----------
gapped_seq1: list
The first gapped sequence.
gapped_seq2: list
The second gapped sequence.
Returns:
--------
numpy.ndarray: A 2D numpy array with two columns containing signal
scores extracted from `gapped_seq1` and `gapped_seq2`
respectively.
"""
gapd_s1 = gapped_seq1
gapd_s2 = gapped_seq2
fl_score = np.zeros((len(gapd_s1),2))
for v1 in range(len(gapd_s1)):
fl_score[v1,0] = gapd_s1[v1][1]
fl_score[v1,1] = gapd_s2[v1][1]
return fl_score
score = signal_map(gapd_s1, gapd_s2)
Sequence alignment decorated with MAb recognition regions¶
Now we can plot the sequence alignment using an ArrayPlotter
instance that higlights sequence recognition regions at the positions
of the respective score residue per alignment column.
To easily interpret the intensity-decorated alignment we will add a
colorbar scaled accordingly. The scale matches the transformation
applied to the recognition signal recorded on the score ndarray.
Let’s build a function to create a custom colorbar object. We will specify the dataframes corresponding to the two antigens screened in this example, the colormap, and the transformation to be represented with the colorbar.
fig = plt.figure(figsize=(8.0, 15))
ax = fig.add_subplot(111)
graphics.plot_alignment_array(
ax, alignments[0], fl_score=score, labels=["FCR3", "NF54"],
show_numbers=True, symbols_per_line=80,
show_line_position=True, label_size=10,
number_size=10, symbol_size=6)
# Add the axes where the colorbar will reside:
ax2 = fig.add_axes([0.13, 0.07, 0.8, 0.01])
ax2.set_frame_on(False)
# Access the colormap of the relevant instace of ArrayPlotter:
colormap = graphics.ArrayPlotter(ax2, score).get_cmap()
def draw_colorbar(axes, array1, array2, colormap,
orient=None, title=None):
df1 = array1
df2 = array2
cmp = colormap
ax = axes
orientation = orient
label = title
# custom Formtatter for tick labels on the colorbar
def fmt(x, pos):
a, b = '{:.1e}'.format(x).split('e')
b = int(b)
return r'${}\cdot10^{{{}}}$'.format(a, b)
vmiA = df1['comb_signal'].min()
vmiB = df2['comb_signal'].min()
vmxA = df1['comb_signal'].max()
vmxB = df2['comb_signal'].max()
# The normalization of this colormap needs to be consistent with the
# data trasnformtion used earlier on this example. The "cubic" law:
norm = mpl.colors.PowerNorm(gamma=0.33, vmin=min(vmiA,vmiB),
vmax=max(vmxA,vmxB))
fig = mpl.pyplot.figure()
return fig.colorbar(mpl.cm.ScalarMappable(norm=norm, cmap=cmp),
cax=ax, orientation=orientation, label=label,
format=mpl.ticker.FuncFormatter(fmt))
# Draw the colorbar
cbar = draw_colorbar(ax2, dfa, dfb, colormap, orient='horizontal',
title='Fluorescence Intensity [AU]')
# To improve readability we tilt the ticklabels on the colorbar
labels = cbar.ax.get_xticklabels()
plt.setp(labels, rotation=45, horizontalalignment='center')
plt.show()
References¶
- 1
U. Iyamu, D. F. Vinals, B. Tornyigah, E. Arango, R. Bhat, T. R. Adra, S. Grewal, K. Martin, A. Maestre, M. Overduin, B. Hazes, S. K. Yanow, “A conserved epitope in VAR2CSA is targeted by a cross-reactive antibody originating from Plasmodium vivax Duffy binding protein,” Frontiers in Cellular and Infection Microbiology, vol. 13, 2023.
