Fig. 1

A Overview of data transformation and analyzes used. Raw counts were transformed to represent normalized data per-library-size, per-total-RNA and per-sample-size (tissue mass). Transformed counts were used to identify stable reference genes free from systematic effect and with subsequent ranking by intra-class correlation. Normalization factors comprised of 10 transcripts from each normalization approach was used in differential expression analysis. B Fold-changes of sample references (average of the top ten stable transcripts per normalization mode) ratios with numerators plotted over columns and denominators over rows. Error bars represent 95% CI. C Transcripts identified as differentially up and down-regulated over time (differences from Week 0 to Week 2 and 12 respectively) from generalized linear models with each normalization factor used as a model offset. Percentages represents proportions of all transcripts identified as differentially expressed regardless of normalization approach. Up- and down-regulation determined from false discovery rate-adjusted p values (p < 0.05). Black points represent intersections, e.g., where the same transcript has been identified from in one or more normalization perspective