Abstract Background The manufacturing processes of plasma products include steps that can remove prions. The efficacy of these steps is measured in validation studies using animal brain‐derived prion materials called spikes. Because the nature of the prion agent in blood is not known, the relevance of these spikes, particularly with steps that are based on retention mechanisms such as nanofiltration, is important to investigate. Study Design and Methods The aggregation and sizes of PrP res assemblies of microsomal fractions (MFs) extracted from 263K‐infected hamster brains were analyzed using velocity gradients. The separated gradient fractions were either inoculated to Tg7 mice expressing hamster‐PrP c to measure infectivity or used in Protein Misfolding Cyclic Amplification for measuring seeding activity. The collected data allowed for reanalyzing results from previous nanofiltration validation studies. Results A significant portion of MFs was found to be composed of small PrP res assemblies, estimated to have a size ≤24 mers (~22–528 kDa), and to contain a minimum of 20% of total prion infectivity. With this data we could calculate reductions of 4.10 log (15 N), 2.53 log (35 N), and 1.77 log (35 N) from validation studies specifically for these small PrP res objects. Conclusion Our gradient data provided evidence that nanofilters can remove the majority of the smallest PrP res entities within microsomes spikes, estimated to be in a size below 24 mers, giving insight about the fact that, in our conditions, size exclusion may not be the only mechanism for retention nanofiltration.