Ugly curves when low %infectec at final dilution (2560). Try more stringent constraints in curve fitting.

[Swarchal]

There's a few examples when the highest dilution (2560) is close to 50% inhibition, that the curve fitting doesn't look great.

e.g:

We are thinking that the IC50 should be shifted more towards the final concentration, so it might be better with a shallower hillslope and/or a higher param_top.

Obviously need to be careful if changing the curve fitting parameters that it doesn't drastically alter the IC50 values of any historical data.