The present work investigates computationally the impact of blending ammonia with ethylene on soot formation in four laminar diffusion flames and the corresponding turbulent partially premixed flames. The conditions conform to experimental investigations (Bennett et al., Combust. Flame, 2020, and Boyette et al., Combust. Flame, 2021) with both sets of flames featuring identical fuel blends. A mass and number density preserving sectional method was applied to all cases with the soot surface growth and oxidation models updated to include reactions with oxides of nitrogen. The gas phase chemistry was extended to include a comprehensively validated ammonia sub-mechanism. For the laminar flames, the inception of soot particles is based on detailed chemistry with pyrene treated as representative of the PAH pool. The accuracy of fitting a global acetylene-based inception step is evaluated for the different fuel blends and subsequently applied to the turbulent flames using a fully coupled transported joint probability density function method featuring an 84-dimensional joint-scalar space. Results obtained for laminar flames show that detailed chemistry coupled with the sectional soot model provides excellent agreement with the measured suppression of the soot volume fraction with increased use of ammonia. Computed particle size distributions (PSDs) show an increase in smaller soot particles under such conditions, consistent with experimental observations. The experimentally observed reduced impact in turbulent flames is also reproduced computationally. The suppression of soot is principally caused by changes in the radical pool leading to reduced soot surface growth and, to a lesser extent, soot inception. The contribution of oxides of nitrogen to soot oxidation is modest. Computed PSDs in laminar and turbulent flames highlight the importance of differences in flame structures and flowfield timescales.