Parent-of-origin effects (POEs) occur when the impact of a genetic variant depends on its parental origin. Traditionally linked to genomic imprinting, these effects are believed to have evolved from parental conflict over resource allocation to offspring, which results in opposing parental genetic influences. Despite their potential importance, POEs remain heavily understudied in complex traits, largely due to the lack of parental genomes.Here, we present a multi-step approach to infer the parent-of-origin of alleles without parental genomes, leveraging inter-chromosomal phasing, mitochondrial and chromosome X data, and sibling-based crossover inference. Applied to the UK Biobank (discovery cohort) and Estonian Biobank (replication cohort), this scalable approach enabled parent-of-origin inference for up to 221,062 individuals, representing the largest dataset of its kind.GWAS scans for more than 60 complex traits and over 2,400 protein levels contrasting maternal and paternal effects identified over 30 novel POEs and confirmed more than 50% of testable known associations. Notably, approximately half of our POEs exhibited a bipolar pattern, where maternal and paternal alleles exert conflicting effects. These effects were particularly prevalent for traits related to growth (e.g., IGF-1, height, fat-free mass) and metabolism (e.g., type 2 diabetes, triglycerides, glucose). Replication in the Estonian Biobank validated over 70% of testable associations.Overall, our findings shed new light on the influence of POEs on diverse complex traits and align with the parental conflict hypothesis, providing compelling evidence for this understudied evolutionary phenomenon.