|NANDOLO, WILSON - University Of Natural Resources & Applied Life Sciences - Austria|
|UTSUNOMIYA, YURI - Sao Paulo State University (UNESP)|
|MESZAROS, GABOR - University Of Natural Resources & Applied Life Sciences - Austria|
|WURZINGER, MARIA - University Of Natural Resources & Applied Life Sciences - Austria|
|KHAYADZADEH, NEGAR - University Of Natural Resources & Applied Life Sciences - Austria|
|TORRECILHA, RAFAELA - University Of Natural Resources & Applied Life Sciences - Austria|
|MULINDWA, HENRY - National Livestock Research Institute|
|GONDWE, TIMOTHY - Lilongwe University Of Agriculture And Natural Resources|
|WALDMANN, PATRIK - Swedish University Of Agricultural Sciences|
|FERENCAKOVIC, MAJA - University Of Zagreb|
|GARCIA, JOSE - Sao Paulo State University (UNESP)|
|Van Tassell, Curtis - Curt|
|CURIK, INO - University Of Zagreb|
|SOLKNER, JOHANN - University Of Natural Resources & Applied Life Sciences - Austria|
Submitted to: Genetic Selection Evolution
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/30/2018
Publication Date: 8/22/2018
Citation: Nandolo, W., Utsunomiya, Y.T., Meszaros, G., Wurzinger, M., Khayadzadeh, N., Torrecilha, R.B., Mulindwa, H., Gondwe, T., Waldmann, P., Ferencakovic, M., Garcia, J.F., Rosen, B.D., Bickhart, D.M., Van Tassell, C.P., Curik, I., Solkner, J. 2018. Misidentification of runs of homozygosity islands in cattle caused by interference with copy number variation or large intermarker distances. Genetic Selection Evolution. 50(1):43. https://doi.org/10.1186/s12711-018-0414-x.
Interpretive Summary: Run of homozygosity (ROH) islands are continuous stretches of homozygous base pairs in the genome occurring in a large proportion of individuals in a population. ROH detection algorithms are based on the degree of similarity of haplotypes on homologous chromosomes in a diploid individual. Coverage gaps and copy number variants (CNVs) may lead to an increase in such similarity, leading to detection of ROHs islands where none exist. This study aimed at identifying ROH islands that may be influenced by marker coverage gaps and/or CNVs, using Illumina BovineHD BeadChip (777 K) single nucleotide polymorphism (SNP) data for Austrian Brown Swiss, Tyrol Grey and Pinzgauer cattle. ROHs were detected using a clustering algorithm implemented in the Golden Helix SNP and Variation Suite (SVS) software, and ROH islands were determined based on population inbreeding levels for each marker. CNVs were detected using multivariate copy number analysis method (CNAM) and optimal segmenting procedures implemented in SVS and a hidden Markov model implemented in PennCNV software. Intermarker distances (IMD) were computed across the genome and regions having average IMD above 9.2365 kilobase pairs (kb) were considered as having SNP coverage gaps. ROH islands which overlapped with CNV regions (CNVRs) and/or SNP coverage gaps were considered potential artefacts. Permutation tests were used to check whether the overlaps between the CNVRs and the ROH islands were due to chance alone. A total of 13, 20, and 24 ROH islands covering 26.637, 36.486 and 35.847 Mb were identified in Brown Swiss, Tyrol Grey and Pinzgauer, respectively. PennCNV detected 306, 613 and 598 CNVRs while SVS detected 187, 150 and 178 CNVRs in the three breeds, respectively. The numbers of consensus CNVRs (CNVRs derived from CNVs detected by both algorithms) were 30, 50 and 75, respectively. Overlaps between ROH islands, CNVRs and/or coverage gaps occurred in 7, 14 and 16 ROH islands, respectively. The intersections between ROH islands and CNVRs were small, but significantly higher compared to ROH islands on random locations across the genome, implying that there was association between ROH islands and CNVRs. This study has shown that a large proportion of ROH islands in Brown Swiss, Tyrol Grey and Pinzgauer cattle may not be real ROH but artefacts of CNVs and/or SNP coverage gaps. This has the implication of potential false positives in studies of selection signatures. To obtain recise identification of ROH islands one needs to account for interference with CNVs or large IMD.
Technical Abstract: Large regions of homozygosity -- or regions where both chromosomes contain the same genetic sequence -- represent a major issue within the field of animal breeding. If an animal contains many regions of homozygosity, this may allow detrimental recessive genes to act to decrease animal performance or to cause a disease state. This study improves the detection of regions of homozygosity by identifying and excluding known artifact regions from analysis, thereby giving a more accurate tool for assessing animal performance loss due to inbreeding.