In another discussion, Terry brought up this recent paper on O2b and Korean patrilineages (with important references to the rest of East Asia) that I believe is worth mentioning:
Soon-Hee Kim et al., High frequencies of Y-chromosome haplogroup O2b-SRY465 lineages in Korea: a genetic perspective on the peopling of Korea. BMC Investigative Genetics 2011. Open access.
Much of the paper is tainted by insistence on lineage age speculations (TMRCA) which help nothing, are unscientific and induce confusion. However I do find interesting that they seem to find (figure 3) that SE Asian instances of O2b (O2b* and O2b1) are derived and not ancestral in relation to NE Asian ones. Hence the occasional O2b found in SE Asia (Indonesia, Vietnam, Thailand, etc.) looks like having originated in Korea or nearby.
On the other hand the haplotype neighbor-joining structures for O3 and C still look like rooted in the South. In relation to NE Asian C (C1 and C3) I find most curious that a single Korean haplotype would seem to hold together C1 and C3, maybe indicating a joint origin within C for these two NE Asian haplogroups.
Overall I get the impression of a South to North colonization in East Asia, which pretty much set the fundamentals of modern population genetics in the whole region. This process was surely followed at a later time by some North to South backflow, which does not need to be "Neolithic" (beginning in South China at least as early as in the North) nor "Han" (in many cases the relevant haplogroups are definitively not found among North Han but Koreans, Japanese, etc.), it may well be, at least partly, related to the extension of blade industries since c. 20 Ka ago, phenomenon that might have got a southwards pointing vector.
In any case, an important reference for data miners.