Manual integration (and peak editing and RT-syncing) use a method where new peaks are created for specified RT range. These peaks were being assigned many of the original attributes that they naturally would have if they were created automatically by peak detection on an EIC. Except for peak rank - which also turns out to be very important for quality scoring. Because of this peaks that are drawn out manually or re-integrated had on average lower quality than those detected automatically, even for exact same m/z and RT. This has been fixed.
Manual integration (and peak editing and RT-syncing) use a method where new peaks are created for specified RT range. These peaks were being assigned many of the original attributes that they naturally would have if they were created automatically by peak detection on an EIC. Except for peak rank - which also turns out to be very important for quality scoring. Because of this peaks that are drawn out manually or re-integrated had on average lower quality than those detected automatically, even for exact same m/z and RT. This has been fixed.