Industrial fermentation processes operate under well defined operating cond
itions to attempt to minimise production variability. Variability occurs fo
r many reasons but a long held belief is that variation in the state of the
seed is highly influential. In this paper a seed stage (a batch process) o
f an industrial antibiotic fermentation is considered and the performance o
f the main production fermentations is correlated with the quality of the s
eed using an unsupervised Kohonen self-organising feature map (SOM). It is
shown that using only seed information poor performance in the final stage
fermentations can be predicted. Data from industrial penicillin G fermenter
s is used to demonstrate the procedure. (C) 1999 John Wiley & Sons, Inc.