Huge discrepancy in the numbers in "Modeling ChIP Sequencing in Silico with Applications"

In your article "Modeling ChIP Sequencing in Silico with Applications", you mentioned that the initial 2,915,382 sequence reads obtained in Robertson's experiments, but when I refer this number to the original paper, the total sequenced reads is 24.1M, which is significantlydifferent from your data. Could you clarify, please?

For our ChIP-seq analysis, we used whatever read sequences that Robertson et al sent to us upon our request, which was made well before the publication of their paper.