Responding to commentary (1), Brettel and Byrdin state (2) that nonproductive back electron transfer (ET) should be only moderately slower than forward ET, instead of the 2.4 ns reported by Liu et al. (3). We believe the authors base their conclusion on erroneous assumptions due to suboptimal experimental design and poor resolution data, as we explain below.First, the quantum yield of 0.5 reported by the Brettel and Byrdin group is too low. The quantum yield of photolyase is in the range of 0.7-0.9, as measured in vivo (4, 5) and in vitro (5, 6) and under single turnover (5) and steady-state (6, 7) reaction conditions by three different groups. An inspection of the supplementary information in the Thiagarajan et al. article (8) shows that only two data points were used for determining substrate saturation. A systematic concentration titration is needed to determine the accurate binding constant. In contrast, the Liu et al. study (3) measured the binding constant with each substrate using multiple data points and knew exactly how much the complex concentration was under each experimental condition.Second, Brettel and Byrdin argue that a rough estimation of the repair quantum yield can be obtained from the absorption (8) used different thymine substrates, and the absorption coefficient at 266 nm could be different, which is directly related to the final absorption value. Thus, the comparison is not valid.Last, Brettel and Byrdin state that the nonproductive back ET should be approximately 400 ps from the double-decay fitting of neutral semiquinone FADH dynamics. Actually, from the "minimal scheme" in the Thiagarajan et al. article (8), the back ET is parallel to the bond splitting. So, only the overall rate of these two steps (the sum of two rates of back ET and bond splitting) can be observed. Given the splitting time of CPD within 90 ps precisely measured in ref. 3, there is no way to obtain a 400-ps decay for FADH (it is also impossible if we take one step back to use the splitting time of 350 ps in their article). In fact, the decay of FADH is the electron return after the CPD splitting. The double exponential decay is from the modulation of active-site solvation.In conclusion, both the Thiagarajan et al. article (8) and a related article in Biochemistry (9) by the same group involve misinterpretation of the steady-state data and low resolution of the dynamics data. Thus, their arguments regarding the shorter unproductive back ET time are not valid.