It was a completely normal "downwind fullstop" call, and according to the available ADS-B data, it appears the pilot initiated the turn on to base at the expected point.
Except from the track indications there is no evidence to say they were commencing a turn, that's all pure speculation. (The curve at the end of the tracking is whilst the aircraft is in rapid descent).
All we know is that they made a normal downwind call, after that the downwind leg appeared normal, and something went terribly wrong in the place we would be expecting them to be thinking about turning base.
I'll add that the plots indicate over 1 minute from the start of downwind to where things seem to have departed normal, a lot can happen in 1 minute of aviation, or nothing may have happened until the split second it departed controlled flight.
I still struggle to see a link to seat rail failure unless the trim was extremely nose down and the pilot let go of the controls. Even if there was a sudden seat rail failure, on downwind in level flight, its hardly going to propel the seat backwards in the same manner as just after lift off whilst accelerating in the climb. But who knows, anything is possible.