I would like to see their comparison to WGS. Frankly, I am not sure how much this strategy helps. Repeats present genome-wide are also repetitive in each chromosome, which still prevents us from getting long contigs. Assembling per chromosome does help to reduce memory, but I do not see this is a major problem nowadays. Flow sorting chromosomes and then constructing barcode libraries add efforts and costs. The contamination rate from other chromosomes is like 10-20% as I remember. These efforts seem not paid off well.

Anyway just my guess. Could be wrong.

This approach is being used to sequence the wheat genome. See the Vitulo et al. 2011 paper in PLoS One, for example.

Major hurdles:

• you need to be able to get the right kinds of cells (metaphase?) in sufficient quantities that let you separate individual chromosomes
• you need to be able to identify chromosomes (size, fluorescence, probes?)
• you need to care (why does it matter?)
• you need to have the relevant manpower (previously genome projects were run by molecular biologists; emerging model organisms sequencing is often run by evolutionary biologists or ecologists who haven't ever even seen a PCR machine)

I think the general attitude is "we'll get what we can with the current technology"... because 2 more years of improvements in sequencing technology will make things automagically better.