Contents

Plot drug resistance frequencies

Plot drug resistance frequencies

Introduction

This notebook creates a table summarising the number of samples with inferred antimalarial drug resistance (DR) in different populations using the data from Plasmodium falciparum version 8 (Pf8) project.

This notebook takes 1 minute to run.

Setup

Install and import the malariagen Python package:

!pip install malariagen_data -q --no-warn-conflicts
import malariagen_data

  Installing build dependencies ... ?25l?25hdone
  Getting requirements to build wheel ... ?25l?25hdone
  Preparing metadata (pyproject.toml) ... ?25l?25hdone
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 4.0/4.0 MB 19.3 MB/s eta 0:00:00
?25h  Preparing metadata (setup.py) ... ?25l?25hdone
  Preparing metadata (setup.py) ... ?25l?25hdone
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 71.7/71.7 kB 5.2 MB/s eta 0:00:00
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 775.9/775.9 kB 42.0 MB/s eta 0:00:00
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 25.9/25.9 MB 66.6 MB/s eta 0:00:00
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 8.7/8.7 MB 94.4 MB/s eta 0:00:00
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 210.6/210.6 kB 15.6 MB/s eta 0:00:00
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 6.3/6.3 MB 94.0 MB/s eta 0:00:00
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 3.3/3.3 MB 81.1 MB/s eta 0:00:00
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 7.8/7.8 MB 82.8 MB/s eta 0:00:00
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 78.3/78.3 kB 5.7 MB/s eta 0:00:00
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 101.7/101.7 kB 7.1 MB/s eta 0:00:00
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 8.9/8.9 MB 91.6 MB/s eta 0:00:00
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 228.0/228.0 kB 16.5 MB/s eta 0:00:00
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 13.4/13.4 MB 89.3 MB/s eta 0:00:00
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 1.6/1.6 MB 60.8 MB/s eta 0:00:00
?25h  Building wheel for malariagen_data (pyproject.toml) ... ?25l?25hdone
  Building wheel for dash-cytoscape (setup.py) ... ?25l?25hdone
  Building wheel for asciitree (setup.py) ... ?25l?25hdone

Import required python libraries that are installed at colab by default.

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import collections
import re
import scipy

Access Pf8 Data

We use the malariagen data package to load the release data.

release_data = malariagen_data.Pf8()
sample_metadata = release_data.sample_metadata()

# take a glance at the metadata dataframe
sample_metadata.head(3)
Sample Study Country Admin level 1 Country latitude Country longitude Admin level 1 latitude Admin level 1 longitude Year ENA All samples same case Population % callable QC pass Exclusion reason Sample type Sample was in Pf7
0 FP0008-C 1147-PF-MR-CONWAY Mauritania Hodh el Gharbi 20.265149 -10.337093 16.565426 -9.832345 2014.0 ERR1081237 FP0008-C AF-W 82.48 True Analysis_set gDNA True
1 FP0009-C 1147-PF-MR-CONWAY Mauritania Hodh el Gharbi 20.265149 -10.337093 16.565426 -9.832345 2014.0 ERR1081238 FP0009-C AF-W 88.95 True Analysis_set gDNA True
2 FP0010-CW 1147-PF-MR-CONWAY Mauritania Hodh el Gharbi 20.265149 -10.337093 16.565426 -9.832345 2014.0 ERR2889621 FP0010-CW AF-W 87.01 True Analysis_set sWGA True

We will only examine the QC pass samples in the analysis of these notebooks.

# Retain QC pass samples
qc_sample_metadata = sample_metadata.loc[sample_metadata['QC pass']]

Access DR Classification and Genotype Data

We can access inferred resistance status classifications of QC-pass Pf8 samples from the Sanger’s cloud storage.

This dataset includes the samples that are predicted to be resistant to 10 drugs or combinations of drugs and to rapid diagnostic tests (RDT) detection: chloroquine, pyrimethamine, sulfadoxine, mefloquine, artemisinin, piperaquine, sulfadoxine- pyrimethamine for treatment of uncomplicated malaria, sulfadoxine- pyrimethamine for intermittent preventive treatment in pregnancy, artesunate-mefloquine, dihydroartemisinin-piperaquine, hrp2 and hrp3 gene deletions.

# Read the data
resistance_classification_fn = pd.read_csv('https://pf8-release.cog.sanger.ac.uk/Pf8_inferred_resistance_status_classification.tsv', sep='\t')

# Rename the first column as 'Sample'
resistance_classification_fn = resistance_classification_fn.rename(columns={resistance_classification_fn.columns[0]: 'Sample'})

# Print the first rows
resistance_classification_fn.head()
Sample Chloroquine Pyrimethamine Sulfadoxine Mefloquine Artemisinin Piperaquine SP (uncomplicated) SP (IPTp) AS-MQ DHA-PPQ
0 FP0008-C Undetermined Undetermined Undetermined Sensitive Sensitive Sensitive Sensitive Sensitive Sensitive Sensitive
1 FP0009-C Resistant Resistant Sensitive Sensitive Sensitive Sensitive Resistant Sensitive Sensitive Sensitive
2 FP0010-CW Undetermined Resistant Resistant Sensitive Sensitive Sensitive Resistant Sensitive Sensitive Sensitive
3 FP0011-CW Undetermined Resistant Undetermined Sensitive Sensitive Sensitive Resistant Sensitive Sensitive Sensitive
4 FP0012-CW Resistant Resistant Sensitive Sensitive Sensitive Sensitive Resistant Sensitive Sensitive Sensitive

We will also use the genotypes utilised in drug resistance status classification containing amino acid and copy number genotypes at six loci: crt, dhfr, dhps, mdr1, kelch13, plasmepsin 2-3. This dataset is also available from the Sanger’s cloud storage.

# Read the data
drm_calls_fn = pd.read_csv('https://pf8-release.cog.sanger.ac.uk/Pf8_drug_resistance_marker_genotypes.tsv', sep='\t', )

# Rename the first column as 'Sample'
drm_calls_fn = drm_calls_fn.rename(columns={drm_calls_fn.columns[0]: 'Sample'})

# Print the first rows
drm_calls_fn.head()
Sample crt_72[C] crt_74[M] crt_75[N] crt_76[K] crt_72-76[CVMNK] crt_93[T] crt_97[H] crt_218[I] crt_220[A] ... mdr1_1034[S] mdr1_1042[N] mdr1_1226[F] mdr1_1246[D] arps10_127-128[VD] fd_193[D] mdr2_484[T] kelch13_349-726_ns_changes mdr1_dup_call pm2_dup_call
0 FP0008-C C I,M E,N T,K CVIET,CVMNK T H I S,A ... S N F D VD D T NaN 0 0
1 FP0009-C C I E T CVIET T H I S ... S N F Y VD D T NaN 0 0
2 FP0010-CW C I,M E,N T,K CVIET,CVMNK T H I S,A ... S N F D VD D T NaN 0 0
3 FP0011-CW C I,M E,N T,K CVIET,CVMNK T H I S,A ... S N F D VD D T NaN 0 0
4 FP0012-CW C I E T CVIET T H I S ... S N F D VD D T NaN 0 0

5 rows × 40 columns

Load all data into single DataFrame

We can merge these 3 datasets (metadata, drug resistance genotype and classification) to facilitate streamlined analysis.

pd.merge(qc_sample_metadata,drm_calls_fn,on='Sample')
Sample Study Country Admin level 1 Country latitude Country longitude Admin level 1 latitude Admin level 1 longitude Year ENA ... mdr1_1034[S] mdr1_1042[N] mdr1_1226[F] mdr1_1246[D] arps10_127-128[VD] fd_193[D] mdr2_484[T] kelch13_349-726_ns_changes mdr1_dup_call pm2_dup_call
0 FP0008-C 1147-PF-MR-CONWAY Mauritania Hodh el Gharbi 20.265149 -10.337093 16.565426 -9.832345 2014.0 ERR1081237 ... S N F D VD D T NaN 0 0
1 FP0009-C 1147-PF-MR-CONWAY Mauritania Hodh el Gharbi 20.265149 -10.337093 16.565426 -9.832345 2014.0 ERR1081238 ... S N F Y VD D T NaN 0 0
2 FP0010-CW 1147-PF-MR-CONWAY Mauritania Hodh el Gharbi 20.265149 -10.337093 16.565426 -9.832345 2014.0 ERR2889621 ... S N F D VD D T NaN 0 0
3 FP0011-CW 1147-PF-MR-CONWAY Mauritania Hodh el Gharbi 20.265149 -10.337093 16.565426 -9.832345 2014.0 ERR2889624 ... S N F D VD D T NaN 0 0
4 FP0012-CW 1147-PF-MR-CONWAY Mauritania Hodh el Gharbi 20.265149 -10.337093 16.565426 -9.832345 2014.0 ERR2889627 ... S N F D VD D T NaN 0 0
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
24404 SPT92049 1306-PF-NG-NGWA-SM Nigeria Oyo 9.592268 8.097575 8.143664 3.618846 2019.0 ERR10893258 ... S N F - VD D T - -1 -1
24405 SPT92054 1306-PF-NG-NGWA-SM Nigeria Oyo 9.592268 8.097575 8.143664 3.618846 2019.0 ERR10893326 ... S N F D VD D T NaN 0 -1
24406 SPT92057 1306-PF-NG-NGWA-SM Nigeria Oyo 9.592268 8.097575 8.143664 3.618846 2019.0 ERR10893333 ... S N F D VD D T - -1 -1
24407 SPT94772 1268-PF-MULTI-PAMGEN-SM Gambia Western 13.451482 -15.372910 13.245396 -16.401559 2017.0 ERR10789456 ... - - - - VD D - - -1 -1
24408 SPT94773 1268-PF-MULTI-PAMGEN-SM Gambia Western 13.451482 -15.372910 13.245396 -16.401559 2017.0 ERR10789457 ... S N F D VD D T NaN 0 0

24409 rows × 56 columns

# Merge the dataframes on "Sample" column
df_all_sample_metadata = pd.merge(pd.merge(qc_sample_metadata,drm_calls_fn,on='Sample'),resistance_classification_fn,on='Sample')

# Print first 3 rows
df_all_sample_metadata.head(3)
Sample Study Country Admin level 1 Country latitude Country longitude Admin level 1 latitude Admin level 1 longitude Year ENA ... Chloroquine Pyrimethamine Sulfadoxine Mefloquine Artemisinin Piperaquine SP (uncomplicated) SP (IPTp) AS-MQ DHA-PPQ
0 FP0008-C 1147-PF-MR-CONWAY Mauritania Hodh el Gharbi 20.265149 -10.337093 16.565426 -9.832345 2014.0 ERR1081237 ... Undetermined Undetermined Undetermined Sensitive Sensitive Sensitive Sensitive Sensitive Sensitive Sensitive
1 FP0009-C 1147-PF-MR-CONWAY Mauritania Hodh el Gharbi 20.265149 -10.337093 16.565426 -9.832345 2014.0 ERR1081238 ... Resistant Resistant Sensitive Sensitive Sensitive Sensitive Resistant Sensitive Sensitive Sensitive
2 FP0010-CW 1147-PF-MR-CONWAY Mauritania Hodh el Gharbi 20.265149 -10.337093 16.565426 -9.832345 2014.0 ERR2889621 ... Undetermined Resistant Resistant Sensitive Sensitive Sensitive Resistant Sensitive Sensitive Sensitive

3 rows × 66 columns

Sulfadoxine-Pyrimethamine (SP) is used for the treatment of uncomplicated cases and Intermittent Preventive Treatment in Pregnancy (IPTp). To make these two status clear in the dataset, we rename the columns.

# Define a dictionary of old&new column names
column_name_changes = collections.OrderedDict()
column_name_changes['SP (uncomplicated)'] = 'SP (treatment)'
column_name_changes['SP-IPTp'] = 'SP (IPTp)'

# Rename the columns using the dictionary
df_all_sample_metadata.rename(columns=column_name_changes, inplace=True)

Table setups

We proceed by defining a dictionary where each key represents a drug along with its associated gene, mutation, column name in the dataset, and sensitivity/resistance indicators.

For details on the heuristics employed to map genetic markers to resistance status classification, please refer to here.

drm_dict = collections.OrderedDict()

drm_dict['Chloroquine'] = {
    'gene': 'crt',
    'mutation': '76T',
    'column_name': 'crt_76[K]',
    'sensitive': 'K',
    'resistant': 'T',
}

drm_dict['Pyrimethamine'] = {
    'gene': 'dhfr',
    'mutation': '108N',
    'column_name': 'dhfr_108[S]',
    'sensitive': 'S',
    'resistant': 'N',
}

drm_dict['Sulfadoxine'] = {
    'gene': 'dhps',
    'mutation': '437G',
    'column_name': 'dhps_437[G]',
    'sensitive': 'A',
    'resistant': 'G',
}

drm_dict['Mefloquine'] = {
    'gene': 'mdr1',
    'mutation': '2+ copies',
    'column_name': 'dup_mdr1',
    'sensitive': 0,
    'resistant': 1,
}

drm_dict['Artemisinin'] = {
    'gene': 'kelch13',
    'mutation': 'WHO list',
    'column_name': None,
}

drm_dict['Piperaquine'] = {
    'gene': 'plasmepsin 2-3',
    'mutation': '2+ copies',
    'column_name': 'dup_pm2',
    'sensitive': 0,
    'resistant': 1,
}

drm_dict['SP (treatment)'] = {
    'gene': 'dhfr',
    'mutation': 'triple mutant',
    'column_name': None
}

drm_dict['SP (IPTp)'] = {
    'gene': 'dhfr and dhps',
    'mutation': 'sextuple mutant',
    'column_name': None
}

drm_dict['AS-MQ'] = {
    'gene': 'kelch13 and mdr1',
    'mutation': '',
    'column_name': None
}

drm_dict['DHA-PPQ'] = {
    'gene': 'kelch13 and plasmepsin 2-3',
    'mutation': '',
    'column_name': None
}

To elaborate on population names in the final table, we define a dictionary listing ten populations with the following abbreviations:

populations = collections.OrderedDict()
populations['SA']       = "South America"
populations['AF-W']     = "Africa - West"
populations['AF-C']     = "Africa - Central"
populations['AF-NE']    = "Africa - Northeast"
populations['AF-E']     = "Africa - East"
populations['AS-S-E']  = "Asia - South - East"
populations['AS-S-FE']  = "Asia - South - Far East"
populations['AS-SE-W'] = "Asia - Southeast - West"
populations['AS-SE-E'] = "Asia - Southeast - East"
populations['OC-NG']    = "Oceania - New Guinea"

Summary of DR sample sizes across populations

We want to create a table summarizing the number of samples by population.

To do this, we will write a function that counts number of samples that are resistant and sensitive for a given drug.

def n_agg(x):
    """
    Aggregate function to calculate the count of non-undetermined samples for each drug in a given population.

    Parameters:
    - x: DataFrame

    Returns:
    - pd.Series: Counts of sensitive samples that are resistant to each drug.
    """

    # Initialize an ordered dictionary to store drug names and their corresponding counts
    names = collections.OrderedDict()

    # Iterate through each drug in drm_dict
    for drug in drm_dict:
        # Count the number of non-undetermined samples for the current drug
        n = np.count_nonzero((x[drug] != 'Undetermined'))

        # Store the drug name and its count in the dictionary
        names[drug] = n

    # Return the result as a pandas Series
    return pd.Series(names)

As an addition to table, we would like to include minimum and maximum drug resistance sample size in the population names.

# Calculate counts of non-undetermined samples for each drug in each population
df_drm_n_table = (
    df_all_sample_metadata
    .groupby('Population')
    .apply(n_agg, include_groups=False)
    .rename_axis(None)
    .transpose()
    .loc[:, populations.keys()]
    .reset_index()
)

# Calculate the minimum and maximum sample sizes across populations
min_n = df_drm_n_table.min()
max_n = df_drm_n_table.max()

# Customize the index to include gene and mutation information for each drug
df_drm_n_table.index = ["%s %s" % (drm_dict[drug]['gene'], drm_dict[drug]['mutation']) for drug in drm_dict]
df_drm_n_table.index.names = ['Marker']

# Rename the columns to indicate their association with drug resistance and sample counts
df_drm_n_table.rename(columns={'index': 'Associated with resistance to'}, inplace=True)

# Update column names to include population names and corresponding sample size ranges
for population in populations:
    new_column_name = f'{populations[population]} (n={min_n[population]}-{max_n[population]})'
    df_drm_n_table.rename(columns={population: new_column_name}, inplace=True)

# Display the final drug resistance count table
df_drm_n_table
Associated with resistance to South America (n=177-224) Africa - West (n=6648-8837) Africa - Central (n=870-1184) Africa - Northeast (n=122-204) Africa - East (n=3599-4000) Asia - South - East (n=81-188) Asia - South - Far East (n=1209-1369) Asia - Southeast - West (n=1665-1884) Asia - Southeast - East (n=2910-5798) Oceania - New Guinea (n=293-340)
Marker
crt 76T Chloroquine 223 8284 1039 196 3944 178 1330 1877 5798 331
dhfr 108N Pyrimethamine 217 8035 1184 200 4000 162 1369 1884 5796 332
dhps 437G Sulfadoxine 224 8007 1155 196 3871 179 1296 1884 5728 331
mdr1 2+ copies Mefloquine 177 6936 923 125 3707 93 1241 1739 4966 307
kelch13 WHO list Artemisinin 180 7220 938 151 3712 144 1335 1775 5295 302
plasmepsin 2-3 2+ copies Piperaquine 182 6648 870 122 3599 81 1248 1779 4868 293
dhfr triple mutant SP (treatment) 224 7474 1058 201 3646 186 1209 1837 5689 338
dhfr and dhps sextuple mutant SP (IPTp) 224 8837 1138 204 3980 188 1216 1665 2910 340
kelch13 and mdr1 AS-MQ 199 7846 1059 160 3940 157 1350 1795 5087 320
kelch13 and plasmepsin 2-3 DHA-PPQ 190 7524 981 157 3804 149 1350 1824 4949 313

Table 1. Numbers of samples used to determine proportions in Table 2.

To save the table:

df_drm_n_table.to_excel("DRM_table_sample_numbers.xlsx")

Summary of DR frequencies across populations

Now, we would like to calculate drug resistance proportions in each population.

We can easily adjust the function to count number of samples that are resistant and sensitive for a given drug, and calculate the proportion.

def proportion_agg(x):
    """
    Aggregate function to calculate the proportion of resistance for each drug in a given population.

    Parameters:
    - x: DataFrame

    Returns:
    - pd.Series: Proportions of resistance for each drug.
    """

    # Initialize an ordered dictionary to store drug names and their corresponding proportions
    names = collections.OrderedDict()

    # Iterate through each drug in drm_dict
    for drug in drm_dict:
        # Count the number of non-undetermined samples for the current drug
        n = np.count_nonzero((x[drug] != 'Undetermined'))

        # Check if there are no non-undetermined samples for the drug
        if n == 0:
            proportion = np.nan  # Set proportion to NaN to avoid division by zero
        else:
            # Calculate the proportion of resistant samples for the drug
            proportion = round(np.count_nonzero(
                (x[drug] == 'Resistant')
            ) / np.count_nonzero(
                (x[drug] != 'Undetermined')
            )*100)

        # Store the drug name and its proportion in the dictionary
        names[drug] = proportion

    # Return the result as a pandas Series
    return pd.Series(names)

Let’s apply this function and create a table summarizing drug resistance proportions by population.

# Create a table summarizing drug resistance proportions by population

# Group the DataFrame by 'Population' and apply the 'proportion_agg' function to calculate resistance proportions
df_drm_table = (
    df_all_sample_metadata
    .groupby('Population')
    .apply(proportion_agg, include_groups=False)
    .rename_axis(None)
    .transpose()
    .loc[:, populations.keys()]
    .reset_index()
)

# Customize the index to include gene and mutation information for each drug
df_drm_table.index = ["%s %s" % (drm_dict[drug]['gene'], drm_dict[drug]['mutation']) for drug in drm_dict]
df_drm_table.index.names = ['Marker']

# Rename the columns to indicate their association with drug resistance
df_drm_table.rename(columns={'index': 'Associated with resistance to'}, inplace=True)

# Update column names to include population names and corresponding sample size ranges
for population in populations:
    new_column_name = f'{populations[population]} (n={min_n[population]}-{max_n[population]})'
    df_drm_table.rename(columns={population: new_column_name}, inplace=True)

# Add '%' symbol to all values in the table
df_drm_table = df_drm_table.map(lambda x: f"{int(x)}%" if isinstance(x, (int, float)) else x)

# Display the final drug resistance table
df_drm_table
Associated with resistance to South America (n=177-224) Africa - West (n=6648-8837) Africa - Central (n=870-1184) Africa - Northeast (n=122-204) Africa - East (n=3599-4000) Asia - South - East (n=81-188) Asia - South - Far East (n=1209-1369) Asia - Southeast - West (n=1665-1884) Asia - Southeast - East (n=2910-5798) Oceania - New Guinea (n=293-340)
Marker
crt 76T Chloroquine 100% 26% 35% 48% 14% 33% 94% 99% 96% 96%
dhfr 108N Pyrimethamine 74% 89% 99% 98% 97% 62% 100% 100% 99% 99%
dhps 437G Sulfadoxine 71% 81% 96% 79% 86% 9% 89% 100% 83% 69%
mdr1 2+ copies Mefloquine 0% 0% 0% 0% 0% 0% 0% 30% 4% 1%
kelch13 WHO list Artemisinin 0% 0% 0% 0% 0% 0% 0% 36% 62% 1%
plasmepsin 2-3 2+ copies Piperaquine 0% 0% 0% 0% 0% 0% 0% 0% 43% 0%
dhfr triple mutant SP (treatment) 0% 79% 88% 51% 83% 2% 46% 86% 87% 0%
dhfr and dhps sextuple mutant SP (IPTp) 0% 0% 5% 2% 4% 0% 13% 79% 12% 0%
kelch13 and mdr1 AS-MQ 0% 0% 0% 0% 0% 0% 0% 10% 3% 0%
kelch13 and plasmepsin 2-3 DHA-PPQ 0% 0% 0% 0% 0% 0% 0% 0% 39% 0%

Table 2. Frequency of different sets of polymorphisms associated with drug resistance in samples from different geographical regions. All samples were classified into different types of drug resistance based on published genetic markers, and represent best attempt based on the available data. Each type of resistance was considered to be either present, absent or unknown for a given sample. For each resistance type, the table reports: the genetic markers considered; the drug they are associated with; the proportion of samples in each major sub-population classified as resistant out of the samples where the type was not unknown. The number of samples classified as either resistant or not resistant varies for each type of resistance considered (e.g. due to different levels of genomic accessibility); numbers in brackets report the minimum and maximum number analysed while the exact numbers considered are reported in Table 1 of this notebook. SP: sulfadoxine-pyrimethamine; treatment: SP used for the clinical treatment of uncomplicated malaria; IPTp: SP used for intermittent preventive treatment in pregnancy; AS-MQ: artesunate + mefloquine combination therapy; DHA-PPQ: dihydroartemisinin + piperaquine combination therapy. dhfr triple mutant refers to having all three of 51I, 59R and 108N in dhfr. dhfr and dhps sextuple mutant refers to having all five of 51I, 59R and 108N in dhfr and 437G and 540E in dhps, plus one of dhfr:164L, dhps:581G, dhps:613S or dhps:613T. Full details of the rules used to infer resistance status from genetic markers can be found here.

To save the table:

df_drm_table.to_excel("DRM_table.xlsx")

previous

Pf8 Analysis Guides: Study Evolution Against Malaria Control Measures

next

Explore drug resistance trends in Africa