According to the findings that we have, we can summarize them to three points:
- The increment of the number of loci analyzed has the power to converge resulting blood meal probabilities to their theoretical numbers. 3-4 loci is sufficient in the experiment.
- PCR bad reactions cause higher variance on resulting multiple blood meal probabilities, but don’t significantly change the proportion of blood meals inferred from the maximum number of allele detected.
- With inbreeding factor, resulting one blood meal probability is overestimated and two-blood-meal probability is underestimated.
Those findings might be useful for the calculating sample size in the laboratory side. We are able to sense how many loci are at least required to generate a credible result under certain circumstances (in our case, 3 loci give a good result already). Although we’ve already known that analyzing more loci give better result, extra locus added to PCR amplifying or electrophoresis might just increase unnecessary laboratory cost and improve the result very little. However, if inbreeding factor is high, it will be a different story. Inbreeding factor interferes the balance of genomic frequency and has negative impact on calculating resulting blood meal probabilities. The number of loci analyzed can compensate this disadvantage. If researcher already knows the area where he is going to collect mosquito samples has serious inbreeding issue, in order to predict more accurate multiple blood meal probabilities, we may suggest him take more loci for analysis. For example, if a researcher is going to the area where the subjects’ inbreeding factor is as high as 0.5, according to our simulation result, he might want to take at least 5 loci to analyze after he collects the samples.