Is it possible?

[ozzy]

Let's say you have a bacterial contamination of the water and you do not know what it are these bacteria, and these are culture resistant they couldn’t grow in vitro. How do we do de novo DNA sequence of this bacterium. How much bacteria must at least in water?

[krobison]

1) Suppose you estimate a genome size of 2Mb, which covers a lot of bacteria
2) Multiple that by the average molecular weight of DNA to get the weight of DNA in one cell
3) Divide 1ug by that to get the number of cells needed for 1 ug
4) Given an estimate of the density of your bacterium in the water, you can now estimate the volume of water you'll need to extract bacteria from.
5) Probably multiply that number by 10 or so to account for incomplete recovery, overestimation of the genome size, etc.

For #2, 1 pg of DNA is just under 1000Mb. So for a 2Mb genome, you'll need aboue 5x10^8 bacteria to provide a microgram (check my math!). If a the bug were present at 1000/liter, then you would need around 5000 liters of water (I made that density up from thin air; you might look for some representative values)

There are kits from various vendors to work with much smaller input amounts than 1 ug; this might be a good place to try one!

[chadn737]

I think I would look around at the metagenomics literature, particularly this paper: http://www.sciencemag.org/content/335/6068/587

[ECO]

Moving to general...

[ozzy]

thank you all.