I am assuming cultured scallops are a sub-population selected from the wild. I have used 7 marker loci that were free of nulls and not out of HWE for studying filed collected insects. The challenge is finding markers with sufficient PIC values in the population. For example, about 5 bi-allelic markers (such as SNPs) are required to replace one multi-allelic marker. Similarly, if you have SSRs with low levels of polymorphism, you may want to increase the number of loci to achieve the best results.
The sample size could be determined by simulation (saturation simulation curves). There are software (e.g. free R-modules) available for simulations to determine the appropriate sample size.
Generally, I think a combination of less than 7 markers is small for any analysis on genetic diversity. because you could be unfortunate to sample only the loci that have very little variation between species. the higher the number of SSR loci studied the better for you. a good number will be more than 10 loci. However if you have fewer which are hyper-variable, with a very High PIC value, then you can confidently work with those.... i am not versed with scallops, but for some species, variability is generally low (low allelic richness, low heterozygosities)... check literature on your species to be sure and if that is the case, work with the highest number or markers you can possibly have ( considering cost and time, of-course.
About population size, ideally, obtaining samples from all individuals will give you the genetic diversity you seek to study. but this is not feasible all the time because of the coast and time involved. so I advice you that if you know where (origin) your samples originate from, then try to get a representative sub-sample of your cultured scallops from all origins in the wild you have. that will thus be a sub-sample of the most variability, enabling you to study the genetic diversity.
I do not know about your study system but I was told that 25 loci are enough for my system (flowering plants). As suggested by others, it also depends on the variability of markers. If you do not need to test individual primer and there is already optimized sets of marker, you can choose as much as you want (more is better). You can also minimize your costs by adjusting markers in a multiplex.
You can use tailed primers if the issue is of cost (see Blacket et al. 2012, Molecular Ecology Resources). You can use multiplex manager to group markers in appropriate multiplexes. However, optimizing multiplex is tedious.
Regarding the number of samples, people prefer to have many. I have seen some people who generally use 20-30 samples for plants. I would have increased the number of markers instead of number of samples. In general, I think 10-20 samples should work.
Thank you very much Omaththage P Perera, Magdalene Ngeve and Dilli Rijal. I am taking notes about your recommendations and examples to try to calculate the sample size and the number of microsatellite loci that I need.
could you list the R functions you used to compute saturation simulation curve? I would like to be able to determine the appropriate number of individuals needed in any system. What elements do I need to determine that?