Background: Pre-exposure prophylaxis (PrEP) reduces HIV acquisition. However, for those who acquire HIV while receiving PrEP there is the potential for drug resistance to arise as a result of partial HIV suppression through PrEP medication exposure before HIV seroconversion is detected. The level of resistance selected by PrEP may not be detectable with standard sequencing methods, which only detect resistance present at frequencies above 20% of the viral population. Studies of antiretroviral treatment failure have suggested that resistance as low as 1% may be clinically significant during subsequent antiretroviral use. 454 ultra-deep sequencing can detect resistance at frequencies below 1%. Methodology: In the Partners PrEP Study, a randomized trial of daily oral tenofovir disproxil fumarate (TDF) or tenofovir plus emtricitabine (FTC), plasma samples from the time of HIV seroconversion were tested for predefined mutations that could be selected by TDF (K65R and K70E) and FTC (K65R and M184IV) using 454 ultra-deep sequencing. Samples were considered resistant if the proportion of mutant sequences was significantly above that observed in wild-type genotype-matched controls using Fisher’s exact test with correction for a 5% false-discovery rate. Plasma levels of tenofovir were measured among seroconverters in the active PrEP arms at the time of seroconversion. Fisher’s exact test was used to determine associations with resistance. Results: A total of 121 seroconverters were tested: 38 had received TDF, 25 FTC-TDF, and 58 placebo. PrEP-related resistance mutations (K65R, K70E and/or M184IV) were detected in 9 (7.4%) at levels >1%, 2 of which had previously been detected by standard sequencing. In the TDF arm, 2 of 38 (5.3%, p=0.65 vs placebo) had resistance >1%: 1 had K65R/K70E and 1 had M184IV. In the FTC-TDF arm, resistance was detected in 5 of 25 (20%, p=0.024 vs placebo): 4 had M184IV alone and 1 had M184IV/K65R. Two of the 58 (3.5%) placebo participants had M184IV resistance. Resistance was detected in 3 individuals retrospectively determined to be HIV infected (seronegative but RNA positive) when PrEP was initiated, and in 6 individuals infected after PrEP start. In seroconverters in the active PrEP arms, PrEP use was associated with higher risk of resistance: 6 of 23 (26%) with measurable drug levels versus only 1 of 39 (2.6%) without evidence of PrEP use had resistance >1% (p=0.009). Conclusions: High-sensitivity resistance testing detected more resistance than standard sequencing in persons acquiring HIV who were exposed to PrEP, although resistance still occurred in a minority of subjects. More resistance was selected by FTC than TDF, as evidenced by the prevalence of M184IV compared to K65R.