Nevertheless, like in any validation, there is probably an expectation for statistical rigor. Not in terms of “lots” but in terms of number of individual samples, analysis methods etc.
The requirement for 3 lots in (manufactured) raw materials stems from the assumption that the product is expected to be uniform (which is obviously not the case with biological samples). As we all know, in reality the product is never 100% uniform - there is variation within-specification and there is noise (variation in uncontrolled variables), and the 3 lots requirement is supposed to (minimally) account for such real-life variation.
I am confident that if you analyse 30 individual samples (not from the same source) per treatment, you will be well-positioned for a rigorous statistical analysis (assuming you are analysing variables and not attributes - the latter will require much larger sample sizes). If samples are difficult to obtain, 20 might do; 15 would be marginal. Below 10-15, statistical distribution determinations become questionable; below 6, standard deviation falls and below 3 even the mean may not be seen as representative.
The requirement for 3 lots in (manufactured) raw materials stems from the assumption that the product is expected to be uniform (which is obviously not the case with biological samples). As we all know, in reality the product is never 100% uniform - there is variation within-specification and there is noise (variation in uncontrolled variables), and the 3 lots requirement is supposed to (minimally) account for such real-life variation.
I am confident that if you analyse 30 individual samples (not from the same source) per treatment, you will be well-positioned for a rigorous statistical analysis (assuming you are analysing variables and not attributes - the latter will require much larger sample sizes). If samples are difficult to obtain, 20 might do; 15 would be marginal. Below 10-15, statistical distribution determinations become questionable; below 6, standard deviation falls and below 3 even the mean may not be seen as representative.