Aneuploidy detection from FFPE archived tissues: A computational approach for FFPE-CUTAC data

For more than a century, Formalin-Fixed Paraffin-Embedded (FFPE) sample preparation has been the standard for long-term preservation of biological material, but severe molecular degradation has long been a technical barrier to high-quality genomic analysis. We recently introduced FFPE-Cleavage Under Targeted Accessible Chromatin (FFPE-CUTAC) with an antibody to RNA Polymerase II (RNAPII) as a sensitive, cost-effective alternative for profiling transcription in FFPEs. Beyond transcription, these profiles present an opportunity to infer aneuploidy (whole chromosome arm gain or loss), a key cancer hallmark. However, the sparse nature of FFPE-CUTAC data presents a computational challenge with no existing fit-for-purpose tools. Copy number alterations (CNA) detection methods for whole genome sequencing (WGS) data typically rely on fold change in read depth. However, our analysis shows moderate correlation between FFPE CUTAC and matched WGS read depth across the genome(Pearson correlation coefficient r = 0.525, Spearman correlation coefficient ρ = 0.688). As a result, directly applying standard read-depth-based CNA methods to FFPE-CUTAC achieves a reasonable but suboptimal recovery rate. These findings indicate the need for a tailored approach that explicitly accounts for the sparsity and noise inherent in FFPE-CUTAC data. Our new method estimates a read depth baseline for each genomic bin rather than a single baseline for the entire sample to better capture the normal copy number level. We partition the genome into 1 Mb bins and use GC content as a genomic feature to group these bins. Bins with similar GC content share the same read depth baseline, defined as the mean read count across all bins within the corresponding GC content group. We evaluate the performance of our strategy using 30 meningioma samples against a standard WGS reference data from the same patient cohort. Our method demonstrated a 95.4% overall aneuploidy detection accuracy (1116/1170 chromosome arms) and ensured a low false-positive rate by correctly identifying 98.9% of all intact arms. By comparison, aneuploidy detection from matching FFPE RNA-seq profiling was lower, with an accuracy of 86.32% (1010/1170 chromosome arms). We established a strategy specifically designed for aneuploidy profiling from FFPE-CUTAC data. This method unlocks the ability to sensitively and affordably identify critical chromosome arm variations from the FFPE-CUTAC data, alleviating the need to generate additional costly WGS data.