TABLE 2.

Bayesian information criteria (BIC) values determined by fitting the function created by combining the candidate slope and intercept functions to the training dataset's D_10% values. Each cell corresponds to the BIC value associated with a given slope and intercept function combination. Green cells indicate smaller BIC values (better performing functions) while red cells indicate larger BIC values (poorer performing functions). BIC values for D_5%, D_20%, D_37%, D_50%, and SF_2Gy can be found in Note S2

Bayesian information criterion (BIC) value associated with the function predicting D10%
Intercept
slope	m	$p·LET$	$m+p·LET$	$m+p·LET+q·LE{T}^2$	$q·e^{s·\mathrm{LET}}$	$q·e^{s·\mathrm{LET}}+m$	$q·e^{s·LET}+p·LET+m$
$c·e^{-f·\mathrm{LET}}$	−155.6	−146.6	−153.4	−147.4	−152.3	−146.4	−141.4
$c·e^{-f·\mathrm{LET}}+g$	−149.9	−140.6	−147.4	−141.4	−146.4	−140.4	−135.4
$c·e^{-f·LET-h·LE{T}^2}+g$	−144.1	−142.5	−142.7	−136.7	−142.0	−136.1	−130.7
$(c+h·LET)·e^{-f·LET}+g$	−144.0	−140.8	−142.4	−136.5	−141.5	−135.5	−130.5
$c·ln(LET-h)·e^{-f·LET}+g$	−144.0	−139.6	−142.2	−136.2	−141.2	−135.9	−130.2
${\displaystyle \frac{c}{\mathrm{\Gamma }(f·LET+h+1)}}+g$	−144.1	−141.9	−142.6	−136.6	−141.8	−135.8	−130.7
${\displaystyle \frac{c{\lambda }^{f·(LET-h)}·e^{-k}}{\mathrm{\Gamma }(f·(LET-h)+1)}}+g$	−138.1	−136.4	−136.7	−130.7	−15.2	−130.0	−127.4