Extended Data Fig. 1: Quality control metrics for m⁶A-seq and technical variation controls.

(A) GC bias correction plot showing the offset applied for GC bias across different GC fractions. Each line represents an individual sample. The offset values are centered around zero, with higher corrections applied at extreme GC fractions, indicating the extent of GC-dependent variability across samples. (B) IP efficiency correction. Log-log plot comparing individual log₂ (odds ratio) values to the mean log₂ (odds ratio) across samples. Each line represents a sample, showing a strong linear relationship, indicating consistency in odds ratio estimates across different conditions. The alignment along the diagonal suggests minimal systematic bias in the data. (C) Boxplots showing the percent relative error across different resampling percentages. Each box represents the distribution of relative errors from multiple resampling iterations. Boxes extend from the 25th percentile (Q1) to the 75th percentile (Q3) of the relative errors, with the median indicated by a horizontal line inside each box. Whiskers indicate the minimum and maximum values within 1.5x the interquartile range (IQR) from the box. Higher resampling percentages yield lower relative errors, reflecting greater estimation stability with larger sample sizes. The total sample size is 100 individuals; for example, a resampling percentage of 50% corresponds to 50 individuals. (D) Violin plots showing the variance distribution explained by variables in m⁶A-seq and RNA-seq. Each violin plot illustrates the data’s density distribution, median, and interquartile range. Box plots within violin plots indicate the median (central line), interquartile range (box edges), and whiskers representing the minimum and maximum values within 1.5x IQR from the quartiles. Individual and regional variance contribute to both datasets. Variance fractions were estimated using linear models with fixed effects as implemented in the variancePartition R package. 93 individuals were included in this analysis. (E) Boxplot showing the number of m⁶A sites per individual, normalized by the total read count for that individual, grouped by brain region. Sample sizes are 18 individuals for BA24, C, and H; 19 individuals for BA9; and 20 individuals for T. The red diamonds represent the mean values for each region. The box spans the 25th percentile (Q1) to the 75th percentile (Q3) of the data (interquartile range, IQR), with the median indicated by a horizontal line inside the box. Whiskers extend to the minimum and maximum values within 1.5x the IQR from the box. Points beyond this range are shown as individual outliers (for example, the dot above BA9). Statistical comparison across brain regions was performed using a one-way ANOVA, resulting in a p-value of 0.651, indicating no statistically significant differences between regions. (F) Boxplots showing distribution of m⁶A values across all 93 samples. The box spans the 25th percentile (Q1) to the 75th percentile (Q3) of the data (interquartile range, IQR), with the median indicated by a horizontal line inside the box. Whiskers extend to the minimum and maximum values within 1.5x the IQR from the box.