Depth of coverage (sequencing)... you need to do the experiment itself and count. Some sequences will amplify easier than others and some will be completely missed out.
There are several different ways express coverage.
- the theoretical "fold-coverage" of a experiment: number of reads * read length / target size
- the theoretical or empirical "breadth-of-coverage" of an assembly: assembly size / target size
- the empirical average "depth-of-coverage" of an assembly: number of reads * read length / assembly size
Is 100,000 reads enough to capture alpha diversity? I am afraid I am unable to answer this one, as we have left my general knowledge on NGS and into the statistic and hands on experience required to estimate the complexity of a microbe habitat. I think the number will have to do with how dominated is a habitat by a particular species, and how easily we can extract DNA from different species of cells. Some species will be more resilient to the protocol we use to extract DNA and thus contribute less DNA, independent of their bio mass in the sample.
I have seen papers that say 100,000 reads is better than 20,000 (Obvious isn't it..and didn't help much.). Perhaps an experiment is required, to see how many reads is needed when a species is 0.1%, 0.01%, 0.001% etc of the total microbial population. 100,000 reads seems small to me but sampling microbial populations by NGS isn't my field.
There are 15 million reads, but only 7.5 million DNA fragments being read (each is being read twice). So at 100,000 reads (I guess DNA fragments), you can run 75 samples.
May your PCR products be long, your protocols short and your boss on holiday