|
To overcome these short-comings in the data, BioEcoSS Ltd. used two computer-intensive methods
to provide useful population estimates. Firstly, to identify real alleles from the "PCR noise",
a genetic algorithm was developed to model the clusters of base-pair measurements,
resulting in 94% of all B-P values contributing to the subsequent analysis.
Secondly, pair-wise comparisons, utilising Monte Carlo simulations, were used
to deduce the probability that any two samples came from the same elephant.
In this way, it was possible to extract the maximum information from otherwise
limited data to give estimated ranges of minimum number alive with acceptably narrow confidence limits.
| |