The amount of sequencing data has increased enormously in the last decade. To analyse the data efficiently, it needs to be assembled to genomes or represented in a compact manner. However, current tools for assembly and compaction of sequencing data only output sequences with no estimates of their correctness which severely hampers accurate estimation of the correctness of downstream analysis. We will develop models to estimate the structural correctness of sequence reconstructions. We will provide for each substring of the sequences the probability that it occurs in the underlying sample. We will consider assembling one genome or a mixture of genomes and compacting sequencing data. Our methods will enable assessing the correctness of downstream genomic analysis accurately and to direct validation efforts to uncertain regions of reconstructed sequences.

2025

2029