[mlu]

This is the result what one would expect. Scientists (until recently) have only been able to sequence species which can be cultured in the laboratory (you need massive amounts of DNA for sequencing). But in fact, more than 90 percent of all microbial species cannot be cultured in the lab and hence (until recently) could not be sequenced and stayed unknown. However, in the past few years, "Next next generation sequencing" (that's how I like to call it) techniques emerged and we are now able to sequence nearly everything. The umbrella terms "metagenomics" and "single-cell sequencing" are often used for such new methods and have huuuge potential in many, many fields. Basically, the new methods eliminate the culturing step and instead have novel techniques for amplifying DNA from only a single strand.

[abandonliberty]

If not sequenced, would we not know many of them by traditional identification methods (e.g. staining) ? I'm guessing their sequencing process prevents the concurrent use of these methods, so we can't match up DNA to known bacteria.

As for unculturability, the recent antibiotic discovery that made the news came from learning how to culture soil bacteria. No, we didn't learn what it needs. We just grew it in it's natural environment: dirt. https://news.ycombinator.com/item?id=8852487

Revolutionary. (seriously).

[ac29]

> If not sequenced, would we not know many of them by traditional identification methods (e.g. staining) ?

Yes, and no. Many "traditional" staining methods dont really offer up much information, like a gram-stain (common when learning microbiology, not so common in research) can only divide bacteria into two groups: gram-positive and gram-negative. More to the point, if these bacteria have no known methods of culturing (which was correctly noted at 90%+), you can not get them in a pure culture and can only stain them in mixed groups, which isnt that useful.

That being said, there are some staining-like methods you can use to identify what taxonomic group a given bacteria is from. You can use a fluorescent DNA probe that binds to a specific target region of DNA that is highly conserved in groups of bacteria (the 16S rRNA). It is not 100% accurate, and it requires reference data from known organisms, but it can be a good tool for initial surveys of mixed samples. You can also get some cool looking pictures from it.

>I'm guessing their sequencing process prevents the concurrent use of these methods, so we can't match up DNA to known bacteria.

Nope! The above method is actually used in some single-cell sequencing techniques. The cell is florescently tagged, then you can use a microfluidic device (or other methods) to isolate the cell, extract its DNA, and sequence. It is however difficult to assemble a complete genome from a single cell's DNA.

[quinndupont]

This assessment strikes me as precisely correct. This article gets dangerously close to click-bait academia, since while it (presumably) does good science, the abstract highlights this sensational-sounding result, that the authors must know will get picked up by the public.

That 1/2 of the DNA is of unknown origin does not mean it is somehow really weird or alien (as the public will imagine), rather, it just means we don't know all that much about extant DNA.

[giarc]

I agree. The whole "NYC Subway" could be replaced with any transit vehicle in any town/city and you would get the same results. Nothing special about NYC subway.

[new299]

The article in question used the Illumina HiSeq 2500 for sequencing. Which would normally be described as a "next-generation" or "2nd generation" sequencer.

I would guess that the rapid reduction in sequencing cost, has in part allowed novel sample preps to be developed which have reduced the amount of material required for a sequencing project (such that even single cell sequencing projects are now feasible).

[ChuckMcM]

Not to mention that petri dishes were broken! [1]

[1] http://schaechter.asmblog.org/schaechter/2014/12/the-great-p...

[mathattack]

The title is definitely linkbait!