Not an expert but there are studies going on. The current process seems to be roughly:
>There are two labs, O'Toole explains, that processed my sample. The first was the wet lab, where, through various molecular assaults, DNA was extracted, 95% of which was bacterial. This was then sent to an external company to be sequenced – there were over 30,000 sequences – and then a huge file of data was crunched by what O'Toole called "a bunch of computer nerds who sit around all day generating stats" in the institute's data lab. (https://www.theguardian.com/science/2014/feb/11/gut-biology-...)
so you get a lot of genetic data mixed up from a lot of strains and it's probably hard to make sense of.
There are some commercial services that will check your bacteria out from $90 https://ubiome.com/consumer/explorer/ https://atlasbiomed.com/uk/microbiome
Humans have thousands of strains (types) of bacteria (100 trillion in human body) - some beneficial, some contributing to and causing ilness.
They settle on some equilibrium in each of us - a balance where the population ratios stay relatively constant (guess).
When you wipe out in one go (with antibiotics) a big chunk of this population they will not necessrily come back to the same balance. And this may be the problem because this new equilibrium may be with some illness-contributing bacteria winning too much of the space.
It's not impossible that the best cure for that is to take antibiotics again followed by probiotics to help recover the ones we know are good. This is also tiny part because you can buy couple of dozens of different strains but your body has thousands - so not sure if this alone is enough. Ideally you'd drink your healthy mom's breast milk - but that may not be an option for all of us :)
I wonder why there are no more big-data-like studies on strain population ratios. I'm guessing because the numbers are counted in trillions, small sample (saliva/urine/bowel?) should give quite good picture? Is it expensive to map it? How is it done to detect presence of bacteria? Is each test focusing only on single strain or you can do it in bulk to map it to some kind of histogram?