RT Journal Article SR Electronic T1 Estimating chromosome sizes from karyotype images enables validation of de novo assemblies JF bioRxiv FD Cold Spring Harbor Laboratory SP 2022.05.22.492982 DO 10.1101/2022.05.22.492982 A1 Arne Ludwig A1 Alexandr Dibrov A1 Gene Myers A1 Martin Pippel YR 2022 UL http://biorxiv.org/content/early/2022/05/24/2022.05.22.492982.abstract AB Highly contiguous genome assemblies are essential for genomic research. Chromosome-scale assembly is feasible with the modern sequencing techniques in principle, but in practice, scaffolding errors frequently occur, leading to incorrect number and sizes of chromosomes. Relating the observed chromosome sizes from karyotype images to the generated assembly scaffolds offers a method for detecting these errors.Here, we present KICS, a semi-automated approach for estimating relative chromosome sizes from karyotype images and their subsequent comparison to the corresponding assembly scaffolds. The method relies on threshold-based image segmentation and uses the computed areas of the chromosome-related connected components as a proxy for the actual chromosome size. We demonstrate the validity and practicality of our approach by applying it to karyotype images of humans and various amphibians, birds, fish, insects, mammals, and plants. We found a strong linear relationship between pixel counts and the DNA content of chromosomes. Averaging estimates from eight human karyotype images, KICS predicts most of the chromosome sizes within an error margin of just 6 Mb.Our method provides additional means of validating genome assemblies at low costs. An interactive implementation of KICS is available at https://github.com/mpicbg-csbd/napari-kics.Competing Interest StatementThe authors have declared no competing interest.KICSKaryotype image-based chromosome size estimator