Schmitt et al. (1) raise the concern that circle sequencing (2) may spuriously count information from the same starting molecule multiple times. There are indeed several mechanisms by which multiple final reads produced by the circle-sequencing process could be derived from the same starting molecule, as discussed briefly in the last paragraph of our Results section (2). As Schmitt et al. (1) point out, the extent to which this occurs in practice can be bounded by using the fact that the ends of reads derived from the same starting molecule will be mapped to the exact same coordinates of a reference genome. An upper bound on the number of duplicate reads in a given experiment can therefore be produced by counting the number of such identical “shear points.” In fact, our statement that “greater than 98.8% of the consensus sequences produced in each dataset were derived from unique circular templates” (2) was based on the exact shear-point filtering analysis suggested in Schmitt et al.’s letter (1).

In the case of duplicated reads produced by the third mechanism suggested by Schmitt et al. (1), there is actually an additional informatic check that can be done. The two potential raw sequences produced by Y-adapter–directed amplification of a double-stranded sheared rolling-circle amplification product will begin at the same circular shift into the original fragment sequence. The value of this shift can be therefore be used as an additional “endogenous tag” (3) to detect duplication produced by this mechanism or by any subsequent amplification.

We agree with Schmitt et al.’s (1) observation that if a depth of sequencing is desired that would saturate the ability of endogenous tags to distinguish original molecules, our method could be modified to incorporate labeling of starting molecules with exogenous sequences. The use of distinguishable sequences tags to eliminate unwanted duplication of information introduced by amplification is itself a rich area of research (4, 5) and would be an exciting direction for further development of circle sequencing.