Abstract Pedunculate oak is seen as an alternative to other tree species for reforestation and mitigation programs in the face of climate change. Thus, the identification of suitable, well-adapted seed sources is of high priority. We perform a genetic inventory on 970 trees from 97 locations in Russia, Belarus, Ukraine and Latvia representing the eastern part of the natural species range. All samples were screened with a set of 348 gene markers covering the nuclear (n = 327) and the organelle genomes (chloroplast: n = 14, mitochondria: n = 7) using a genotyping-by-sequencing method. We identified 10 plastidial haplotypes, which showed a strong spatial pattern with a highly significant autocorrelation up to a range of 600 km. After excluding 47 potentially selected markers, significant spatial genetic structure up to 800 km was also observed at the selectively neutral nuclear loci. Bayesian clustering analysis indicated a sub-division into two gene pools present at all locations but with a shift of proportion from West to East. The genetic diversity was negatively correlated with longitude, showing higher values in the West and Centre and lower values at the northern, eastern and southeastern edges of the natural distribution range. We sub-divided the sampled area into six geographic regions and used the genetic composition of all samples in each region as a genetic reference for assignment tests. Self-assignment of groups of 10 individuals to their region of origin showed an average assignment success of 93 per cent. Samples of three artificial plantations of pedunculate oak in Siberia could be clearly assigned to the East–North and Centre–South regions. We discuss the implications of our results for the forest practice and in particular the proofing of seed sources.