Not necessarily.
It is common to see differences between percent 6mA reported (and other QC metrics) from QC sequencing and full sequencing. Reasons include:
The QC dataset is small, around 100 Mb, while full datasets are much larger.
Different sequencing platforms may be used with distinct technologies and algorithms for calling 6mA bases. For example, Plasmidsaurus uses ONT and your institution may use a PacBio sequencer for full sequencing.
Confidence thresholds and other on-device settings may differ between sequencers and runs.
These factors can shift the reported average percent 6mA, which in turn can influence output metrics from fiberseq-qc. However, the change should not be dramatic. More importantly, the biological patterns should remain consistent. Your MSP length distributions, nucleosome size histograms, and autocorrelation plots should look similar between QC and full datasets (see this article). If the overall chromatin patterns are consistent and biologically reasonable (e.g., your nucleosome footprint lengths are around 147bp), your labeling likely worked as intended.