Characterising tandem repeat complexities across long-read sequencing platforms with TREAT

Tandem repeats (TR) play important roles in genomic variation and disease risk in humans. Long-read sequencing allows for the characterisation of TRs, however, the underlying bioinformatics perspective remains challenging. We evaluated potential biases when genotyping >864k TRs using diverse Oxford Nanopore Technology (ONT) and PacBio long-read sequencing technologies. We showed that, in rare cases, long-read sequencing suffers from coverage drops in TRs, such as the disease-associated TRs in ABCA7 and RFC1 genes. Such coverage drops can lead to TR mis-genotyping, hampering accurate assessments of TR alleles and highlighting the need for bioinformatic tools to characterise TRs across different technologies and data-types. For this reason, we have developed otter and TREAT: otter is a fast targeted local assembler, cross-compatible across different sequencing platforms. It is integrated in TREAT, an end-to-end workflow for TR characterisation, visualisation and analysis across multiple genomes. Together, these tools enabled accurate characterisation of >864k TRs in long-read sequencing data from ONT and PacBio technologies, with error rates ranging 0.6-1.1% and with limited computational resources. This performance extends across diverse genomes: applied to clinically relevant TRs, TREAT significantly detected diseased individuals with extreme expansions (p=4.3x10-7 and p=1.4x10-5, TR expansions in RFC1 gene). Importantly, in a case-control setting, we significantly replicated previously reported TRs-associations with Alzheimer's Disease, including those near or within APOC1 (p=2.63x10-9), SPI1 (p=6.5x10-3) and ABCA7 (p=0.04) genes. Our tools overcome common limitations regarding cross-sequencing platform compatibility and allow end-to-end analysis and comparisons of tandem repeats in human genomes, with broad applications in research and clinical fields. ### Competing Interest Statement The authors have declared no competing interest.