This conclusion led by the results is dubious. Original paper is available at http://www.nature.com/srep/2012/120809/srep00570/full/srep00570.html
First, when we see figure 3 on the linked paper, F2 generation in Motomiya indicates extremely highest abnormality rate despite the fact that Motomiya is farther than Hirono from Fukushima Daiichi nuclear power plants ,and mountains lie between the nuclear power plants and Motomiya.
Second, figure 1 and figure 4 are also dubious. The problem found in figure 1d is that three dots locating on larger than 1μSv/h cannot be approximated by linear function. Four dots on larger than 0.4μSv/h on figure 4b seem to have the similar problem.
Surely, we can say that radioactive fallout affected P generation and F1 generation directory. But, it is doubtful that the abnormality genes were trandmitted to offsprings.