RNA fractionation followed by sequencing is a powerful strategy to study RNA localisation, translation, and subcellular compartmentalisation. However, this approach introduces a fundamental compositional challenge: library preparation and sequencing depth often distort the original proportions of RNA between fractions. These distortions can severely bias differential analyses, particularly under biological conditions that shift RNA distribution across compartments or ribosome-associated fractions.
We present FracFixR, a statistical framework designed to recover original RNA fraction proportions and enable accurate comparisons across fractionated RNA-seq datasets. FracFixR models the compositional relationship between total and fractionated RNA libraries using non-negative linear regression on selected transcripts to estimate global fraction weights and the contribution of an unobserved, “lost” fraction. It corrects transcript-level abundances and supports robust differential proportion testing using binomial, beta-binomial, or logit-based models.
FracFixR was rigorously validated on synthetic datasets with known ground truth, as well as real polysome profiling data from different B-cell lines. The method consistently recovered true fraction weights with high accuracy and effectively adjusted transcript abundances (Pearson r > 0.85), even under the conditions of uneven sequencing depth or substantial RNA loss. Application to polysome-associated fractions from acute lymphoblastic leukaemia (B-ALL) subtypes revealed over 1,700 differentially translated transcripts and highlighted translational regulatory differences linked to relapse phenotypes.
FracFixR is implemented as an R package (available on GitHub) and integrates seamlessly into RNA-seq analysis workflows. It provides both compositional normalisation and visualisation tools, allowing users to extract biologically meaningful insights from fractionated RNA-seq data while accounting for technical and compositional biases.
By explicitly modelling sample composition, FracFixR fills a critical gap in the analysis of RNA departmentalisation and translation-state profiling, enabling more accurate interpretation of RNA-seq experiments involving fractionation protocols.