Intrinsically, the periodical release should be made only primary data, so, TPA data should not be included. However, due to some disorders of sorting sequence data, TPA entries in bulk sequence data from ENA were continuously included like as primary ones at DDBJ. The problem was found when we were preparing to change the structure of directories on anonymous FTP site.
In the release, rel. 120 (June 2020), we fixed the problem, so, the number of records looks like decrease from rel. 119. There are mistakes in the statistics of the periodical releases from rel. 113 (June 2018) to rel. 119 (March 2020), and we have corrected the summaries of statistics of DDBJ releases on Statistics site, and we have decided not to make any corrections for the past release notes.
We apologize for any inconvenience caused by this issue.