If we cross Old Blush with a wild collected Rosa multiflora and find that 5 out of 10 seedlings rebloom, we may assume that the Multiflora carries the rebloom trait. However, if the same cross gives one rebloomer out of 150 seedlings, uniparental disomy is probably at work -- non-dysjunction being more common than "mutation".
In Rosa and Rubus, sectional characters are integrated into supergenes. If uniparental disomy occurred in an intersectional cross, the hybrid could lack the specific characters of one parent without precisely resembling the other. We should not be too quick to judge a seedling "purely maternal" where there has been reasonable care in protecting against accidental pollination.
"The frequency with which uniparental disomy occurs is unknown, since if abnormal disease genes are not involved, its occurrence may go undetected. In the case of cystic fibrosis, one estimate is that 1/500 affected individuals may be due to maternal uniparental disomy. The recurrence risk of CF in subsequent offspring would be significantly lower in cases of uniparental disomy (probably <1%) than in the usual situation in which each parent is a heterozygous carrier of the abnormal CF gene, with a recurrence risk of 25% with each pregnancy. Hence, it is wise to do genetic studies on both parents to be sure they are both carriers in order to provide the most accurate recurrence risk information."
"While the mechanism(s) leading to uniparental disomy 15 have not yet been fully delineated, there is evidence for an initial trisomy 15 conception, followed by chromosome loss. Specifically, there have been case reports of trisomy 15 found on chorionic villus sampling at 10-11 weeks gestation, followed by amniocentesis at 16-18 weeks documenting a viable fetus with two 15 chromosomes, and the resulting child having Prader-Willi syndrome on the basis of maternal uniparental disomy. This suggests non-dysjunction as a common mechanism for uniparental disomy which is supported by the observation of an advanced maternal age effect in cases of Prader-Willi syndrome due to maternal disomy, but not for those due to paternal interstitial deletion of 15q."
Full text of: Uniparental Disomy and Genomic Imprinting