© 2018 Elsevier Inc. (Molecular Cell 65, 432–446; February 2, 2017) While doing analysis using our own data, we noticed that the GEO datasets uploaded in connection to the above article had the following errors. For some of the samples, incorrect replicates or incomplete fastq files were selected for GEO submission. In addition, the numbers of raw sequencing reads and the total valid read pairs in some of the samples were inadvertently inflated by counting some reads more than once due to a technical error. These incorrect replicates and redundant counts did not reflect the actual fastq files that were used in the computational analyses, and thus all the results reported in the paper remain unaffected. We have replaced these files in GEO with the correctly counted files and have now updated this article's Table S1, which referred to these GEO datasets. We apologize for any inconvenience.