The skindata csv file contains the maximum permeation rate for each compound in the logJmax column.It also contains a number of physicochemical properties for each compound:

logSaq is the aqueous solubility;

logKow is the octanol-water partition coefficient, measuring lipophilicity;

Mpt is the melting point;

MV is the molecular volume.

1. Calculate regression statistics (R², slope with error, y-intercept with error, residual standard error) for each of these physicochemical properties as an individual predictor, and the maximum permeation rate logJmax as the response. Enter them in the table below. [4 marks]

Property	R2	Slope (m)	Error in slope	y-intercept (c)	Error in y-intercept	Residual standard error
logSaq	0.000220	-0.000314	0.003122	-4.3106	0.5278	1.6831
logKow	0.005869	-0.002029	0.003812	1.2146	0.6635	2.1465
Mpt	0.01915	0.2049	0.2186	346.87	37.688	121.55
MV	0.16750	0.4019	0.1293	85.81	22.511	72.829

1. Which model gives the best prediction for logJmax, based on your findings? Justify your answer. [2 marks]

The logSaq scatter plot is the best predictor for logJmax as it shares the most similar regression line of best fit, as well as the most similar spread of data. Not only this, but the slope, error in slope and the residual standard error of logKow is in close proximity to the slope, error in slope and the residual standard error of logJmax. This can be seen in the table below.

Property	R2	Slope (m)	Error in slope	y-intercept (c)	Error in y-intercept	Residual standard error
logSaq	0.000220	-0.000314	0.003122	-4.3106	0.5278	1.6831
logJmax	0.037700	-0.004354	0.003175	-6.3869	0.5527	1.7880

1. Include a scatter plot of logJmax against the best predictor, including the regression line of best fit. [1 mark]