More rational approaches have been developed which reduce the empirical nature of strain-improvement programmes. These developments include the streamlining of the empirical techniques and the use of more directed selection methods. Before discussing the attempts at directed selection the empirical approach will be considered along with the attempts made to improve this approach, including miniaturized programmes.
The empirical approach to strain improvement involves subjecting a population of the micro-organism to a mutation treatment and then screening a proportion of the survivors of the treatment for improved productivity. The assessment of the chosen survivors was usually carried out in shake flasks, resulting in the procedure being costly, both in terms of time and personnel. According to Fantini (1966) the two questions which arise in the design of such programmes are:
(i) How many colonies from the survivors of a mutation treatment should be isolated for testing?
(ii) Which colonies should be isolated?
In attempting to answer the first question Fantini dismissed statistical approaches as impractical and claimed that the number of colonies isolated for testing is determined by the practical limiting factors of personnel, incubator and shaker space and time. However, Davies (1964) demonstrated that a statistical approach could give valuable guidelines for the efficient utilization of physical resources in strain-development programmes. Davies based a computer simulation of a mutation and screening programme on the availability of 200 shaker spaces and practical results of error variance and the distribution of yield likely to occur amongst the mutants. He assumed that the majority of the progeny of a mutation would give a small, rather than a large, increase in productivity and it would, therefore, be more feasible to screen a small number in the hope of obtaining a small increase rather than screen a large number in the hope of obtaining a large increase. The major difficulty inherent in this approach is the error involved in determining small increases which implies the replication of screening tests and, therefore, the use of more facilities.
Davies used the computer simulation to investigate the merits of replication in screening programmes and finally proposed the use of a two-stage scheme where mutants were tested singly in the first stage and then the better producers were tested in quadruplicate in the second stage. Davies concluded that such two-stage schemes were adequate over a wide range of conditions, although a one-stage screen could be used if the testing error were small and the frequency of occur rence of favourable types high, and a three-stage scree if the testing error were high and the frequency q, occurrence of favourable types low. A screening pro gramme based on Davies' proposals is shown in P; 3.24. ™
The answer to Fantini's second question (which colonies should be isolated?) is extremely difficult in the field of secondary metabolites and in Davies' scheme the colonies were chosen at random. The selection of colonies on the basis of changed morpholow has been considered by a number of workers and it appears that this is an undesirable technique. Elander (1966) demonstrated that it was preferable to isolate normal morphological types as, although a morphologi-
Was this article helpful?