S2 Text.

Defining the in-vivo phenotype of artemisinin resistant
falciparum malaria: A modelling approach
Supporting information 2 - data plotted with predicted composite distributions
from mixture models
A plot of the most parsimonious model (red) as a probability density function (red
line) with a histogram of the clearance half-life data (grey) for the Northwestern ThaiMyanmar border in 2001.
1
A plot of the most parsimonious model (red) as a probability density function (red
line) with a histogram of the clearance half-life data (grey) for Northwestern ThaiMyanmar border in 2002.
2
A plot of the most parsimonious model (red) as a probability density function (red
line) with a histogram of the clearance half-life data (grey) for the Northwestern ThaiMyanmar border in 2003.
3
A plot of the most parsimonious model (red) as a probability density function (red
line) with a histogram of the clearance half-life data (grey) for the Northwestern ThaiMyanmar border in 2004.
4
A plot of the most parsimonious model (red) as a probability density function (red
line) with a histogram of the clearance half-life data (grey) for the Northwestern ThaiMyanmar border in 2005.
5
A plot of the most parsimonious model (red) as a probability density function (red
line) with a histogram of the clearance half-life data (grey) for the Northwestern ThaiMyanmar border in 2006.
6
A plot of the most parsimonious model (red) as a probability density function (red
line) with a histogram of the clearance half-life data (grey) for the Northwestern ThaiMyanmar border in 2007.
7
A plot of the most parsimonious model (red) as a probability density function (red
line) with a histogram of the clearance half-life data (grey) for the Northwestern ThaiMyanmar border in 2008.
8
A plot of the most parsimonious model (red) as a probability density function (red
line) with a histogram of the clearance half-life data (grey) for the Northwestern ThaiMyanmar border in 2009.
9
A plot of the most parsimonious model (red) as a probability density function (red
line) with a histogram of the clearance half-life data (grey) for the Northwestern ThaiMyanmar border in 2010.
10
A plot of the most parsimonious model (red) as a probability density function (red
line) with a histogram of the clearance half-life data (grey) for the Northwestern ThaiMyanmar border in 2011.
11
A plot of the most parsimonious model (red) as a probability density function (red
line) with a histogram of the clearance half-life data (grey) for the Northwestern ThaiMyanmar border in 2012.
12
AIC values for the North-Western Thai-Myanmar border stratified by year
Components
1
2
3
4
5
2001
11.3
12.6
18.0
17.5
17.6
2002
85.6
82.3
89.5
86.0
89.3
2003
66.8
62.3
61.0
63.6
62.7
2004
39.4
37.4
42.5
39.5
43.8
2005
16.6
21.1
22.4
24.0
27.0
2006
9.7
13.3
15.9
20.7
0.0
2007
26.1
30.3
34.5
34.8
40.3
2008
416.4
407.0
411.5
416.0
422.0
2009
316.1
314.4
319.1
319.5
324.5
2010
182.9
175.8
179.2
175.2
179.7
2011
125.0
107.7
110.8
108.2
113.6
2012
78.3
44.5
48.5
44.5
44.7
If the condition of minimum AIC is used rather than the condition for most
parsimonious model given in the methods section of the main text, then the plot of
the means of the best fitting mixture models compared with the violin plots of the
data would be as follows:
13
AIC values for the full dataset stratified by country
Components
1
2
3
4
5
All data
2043.9
1928.3
1932.1
1927.9
1933.3
Northwestern Thai-Myanmar border
1697.1
1608.1
1612.4
1605.3
1607.8
Western Cambodia
139.5
116.0
112.7
111.5
112.3
If the condition of minimum AIC is used rather than the condition for most
parsimonious model given in the methods section of the main text, then the plot of
the means of the best fitting mixture models compared with the violin plots of the
data would be as follows:
14