Abstract
The arrangement of markers on the genome can be defined in either physical or linkage terms. While a physical map represents the inter-marker distances in base pairs, a genetic (or linkage) map pictures the recombination rate between pairs of markers. High-resolution genetic maps are key elements for genomic research, such as fine-mapping of quantitative trait loci, but they are also needed for creating and updating chromosome-level assemblies of whole-genome sequences. Based on published results on a large pedigree of German Holstein cattle and newly obtained results with German/Austrian Fleckvieh cattle, we aim at providing a platform that allows users to interactively explore the bovine genetic and physical map. We developed the R Shiny app CLARITY available online at https://nmelzer.shinyapps.io/clarity and as R package at https://github.com/nmelzer/CLARITY that provides access to the genetic maps built on the Illumina Bovine SNP50 genotyping array with markers ordered according to the physical coordinates of the most recent bovine genome assembly ARS-UCD1.2. The user is able to interconnect the physical and genetic map for a whole chromosome or a specific chromosomal region and can inspect a landscape of recombination hotspots. Moreover, the user can investigate which of the frequently used genetic-map functions locally fits best. We further provide auxiliary information about markers being putatively misplaced in the ARS-UCD1.2 release. The corresponding output tables and figures can be downloaded in various formats. By ongoing data integration from different breeds, the app also facilitates comparison of different genome features, providing a valuable tool for education and research purposes.
1 Introduction
Genomic research involving gene mapping of economically important traits, population-specific genetic structure and evolutionary history relies heavily on genetic maps built on the extent of linkage disequilibrium (LD) between genomic markers (e.g., ; ). For example, to what extent LD persists in a certain genomic region determines the number of markers required to fine-map a quantitative trait loci with succinct power and precision (for review see ). Moreover, genetic maps are valuable resources for comparative genomic analyses among breeds or species (e.g., ; ). Of utmost topical importance, however, is the contribution of genetic maps (also known as linkage maps) to measuring haplotype similarity in the context of genomic selection () and to chromosome-level assemblies of whole-genome sequences (e.g., ; ).
Given the value of cattle in sustaining the world food security, the bovine genome is subject of vast amount of ongoing research. We recently updated the genetic map of German Holstein breed () and compared it with physical coordinates of the most recent bovine reference genome assembly ARS-UCD1.2 (). As an extension to this resource, here we introduce a Shiny app CLARITY which facilitates interactive visualisation of the bovine genetic and physical map. CLARITY illustrates the details of male recombination across the bovine genome of selected breeds and suggests suitable genetic-map functions. In addition to published findings, results have been updated by taking most recent knowledge about putatively misplaced markers in the bovine genome assembly into account (). Moreover, a linkage map for German/Austrian Fleckvieh cattle has been created. The CLARITY app can therefore serve as a toolkit for both educational and research purposes for the genome of bovine and related species.
2 Data preparation
The app provides access to linkage maps built based on the 50K genotypes of a large pedigree of German Holstein cattle (1,053 half-sib families which comprise 367,056 genotyped animals) as well as German/Austrian Fleckvieh cattle (298,850 genotyped animals pedigreed across 6,866 half-sib families) and based on estimates of recombination rate between intra-chromosomal marker pairs. In what follows, we briefly describe the workflow towards genetic coordinates as depicted in Figure 1 according to . As some of the steps require several days of computing or visual inspection, the workflow was executed once in advance; the app itself dynamically processes physical and genetic coordinates as well as pairwise recombination rates. If not stated otherwise, data processing was executed in R v4.1.3 ().
FIGURE 1
Step 1:Genotype data were filtered for minor allele frequency >1% and for Mendelian inconsistencies both on marker and individual level using PLINK v1.9 () with recommended settings. Genotypes with a Mendelian inheritance error were set to “NA” and missing values were imputed using Eagle v2.4.1 program (). Putatively misplaced markers, which have recently been reported (), were discarded. Data passed on to Step 2 comprised 876 half-sib families with 366,565 progeny genotyped at 44,696 single nucleotide polymorphisms (SNPs) in Holstein and 1,577 half-sib families with 270,636 progeny genotyped at 40,003 SNPs in Fleckvieh.
Step 2:Paternal recombination rates to build the genetic map were derived from two different methods, see Figure 1. First, the deterministic approach developed by , which is implemented in the R package hsphase v2.0.2, yielded estimates of recombination rate between adjacent markers. These estimates were later used to form the landscape of recombination hotspots. Genetic coordinates were estimated as cumulative sum of recombination rates between neighbouring markers. Second, the likelihood-based approach, as implemented in the R package hsrecombi v0.3.4 (), was applied to estimate recombination rates between all intra-chromosomal marker pairs. These estimates of recombination rate also enabled the identification of candidate misplaced markers in the current assembly release ARS-UCD1.2 (). Genetic coordinates were obtained with a smoothing approach, in which all recombination rates < 0.05 between any intra-chromosomal marker pairs were taken into account.
Step 3:As a novel contribution, the relationship between recombination rate and genetic distance between all intra-chromosomal marker pairs was investigated for a set of commonly used genetic-map functions (; ; ; ). Given the estimates of recombination rate between two markers i and j and its genetic distances derived in Step 2 (), a genetic-map function was fitted to the data by solving the following minimisation problem in terms of the model parameter a (where p is the total number of markers per chromosome):We solved this optimisation problem using the R function optim with “Brent” option allowing to specify restrictions on a. Rao’s system of mapping function requires . Furthermore, instead of Haldane’s original map function, we investigated a scaled version thereof, i.e.For the Binomial map function of , we employed a grid search over seeking the minimum squared deviation as described above. The fitted function leading to the least squared deviation constituted the “best” genetic-map function.
3 Implementation
The CLARITY app is an R Shiny web GUI for various operating systems. It relies on several R packages to enable the outcome and visualisation functionalities. CLARITY was implemented in R v4.1.3 () with help of the R packages shiny v1.7.1 () and shinydashboard v0.7.2 () to create a dashboard. The graphical output was produced using the R packages ggplot2 v3.3.5 (), plotly v4.10.0 () and ggVennDiagram v1.2.0 (). To provide additional helpful features for the app, such as hiding or toogle of elements, the R packages shinyjs v2.1.0 () and shinycssloaders v1.0.0 () were incorporated. Tables were generated using the R package DT v0.22 (). Further R packages were employed: cachem v1.0.6 (), config v0.3.1 (), dplyr v1.0.10 (), gridExtra v2.3 (), htmltools v0.5.2 (), magrittr v2.0.3 (), metathis v1.1.2 (), rlang v1.0.6 (), sf v1.0.8 (), RVenn v1.1.0 () and purrr v0.3.5 (). Eventually, an R package was built from the CLARITY app with use of the R package golem v0.3.2 (), which offers default R files for creating the package as well as for deploying the app. The R package roxygen2 v7.1.2 () was employed for package documentation. The processed data (i.e., recombination rates, genetic coordinates and parameters of genetic-map functions) were included as Rdata files in the folder “extdata”.
We optimized the app following recommendations for best practice with lighthouse (). Figures were compressed with the tool Squoosh (), and caching of those figures requiring longer loading was enabled.
The structure of the app relies on modules, which eases a clear and concise organisation. The use of modules and corresponding interfaces to the main shiny ui and shiny server enables a straightforward maintenance of the software. Furthermore, since each module is an independent app with its own interface and server (), future modification and extension of the app are supported.
4 Realisation and features
The app has three sidebar menus: “Information”, “Breed analysis” and “Breed comparison” which are described in the following.
4.1 Information
The first sidebar menu comprises two subitems: (1) general information about the project and contact options as well as (2) details of resources used. More specifically, subitem (2) contains a brief data description for each breed and outlines the methodology used for data analysis and parameter estimation. This subitem also provides auxiliary information about candidate markers identified as being putatively misplaced and/or residing in problematic regions of the ARS-UCD1.2 release (, see also Figure 2). These markers were recommended to be excluded from subsequent genomic analyses, such as phasing, imputation or genome-wide association studies.
FIGURE 2
4.2 Breed analysis
Under the second sidebar menu “Breed analysis”, all tabulated and graphical outcomes are presented for the available breeds. So far, options “Holstein” and “Fleckvieh” are available. The user can select a single chromosome or all chromosomes to interconnect the physical and linkage map. The results are divided into different tabs (implemented as separate modules) within the main panel: general information, genetic map, hotspot detection and genetic-map functions. Generally, outputs including tables and figures for a certain interval or for the entire data can also be downloaded for being locally stored. The properties of each tab are explained in more detail below.
4.2.1 General information
For all chromosomes, summary statistics are provided about the number of markers considered, total number of recombination events, length of physical and both genetic maps (from Step 2) in tabulated format. This table reduces if a single chromosome is selected (Figure 3). An interactive graphical output displays physical versus genetic length per chromosome.
FIGURE 3
4.2.2 Genetic map
Physical and genetic map coordinates are listed for all markers in table or graphical format. If a specific chromosome is selected on the sidebar menu, the user can zoom into relevant chromosomal regions (Figure 4). In each graphic, the genetic-map coordinates of the deterministic and likelihood-based approach appear. The graphical as well as the table output is adaptable to a user specified chromosome window. Selecting the option “all chromosomes” provides a static overview of 29 single graphics.
FIGURE 4
4.2.3 Hotspot detection
The CLARITY app offers a landscape of putative recombination hotspots, in which marker intervals with an elevated recombination rate are colour-marked across the bovine genome or a selected chromosome (Figure 5). The default threshold for the recombination rate is adopted from who defined a hotspot region with a recombination rate exceeding 2.5 standard deviations from the genome-wide average. The threshold is adjustable by the user. Changing the threshold accordingly affects the interactive graphic as well as the corresponding table listing all markers within the hotspot intervals.
FIGURE 5
4.2.4 Genetic-map functions
The user can investigate the suitability of frequently used genetic-map functions and their overall and local fit to the observed recombination activity (Figure 6). The parameter specifying a genetic-map function was estimated in Step 3 which took all intra-chromosomal marker pairs into account.
FIGURE 6
The fitted genetic-map functions are illustrated together with a reduced scatterplot of recombination rate versus genetic distance for computational reasons. Especially for chromosomes with 1,415 (i.e., >1,000,000 marker pairs), the computing time is drastically increased retarding the visualisation. To ensure smooth processing, data were thinned based on the Euclidian distance among consecutive data points; a data point is a pair of recombination rate and genetic coordinate ordered according to a vectorised triangular matrix of SNP identifiers. In total, 200,000 data points with largest Euclidian distance were kept. This reduction of data did not impair the visual appearance of the scatter plot.
4.3 Breed comparison
The third sidebar menu “Breed comparison” contains comparative analyses between breeds separated into the same tabs as described above. In addition, a Venn diagram summarises numbers of breed-specific and shared SNPs on a selected chromosome or over the entire autosome as well as in hotspot intervals. In the tabs “genetic map” and “hotspot detection”, the Venn diagram is interactively linked with the corresponding table—this allows the user to retrieve particular information of a selected Venn set. The Venn diagram also dynamically adapts to a user defined range and threshold, respectively. Furthermore, since the fit of genetic-map functions might differ between breeds and chromosomes, a barplot displays counts of “best” genetic-map functions in each breed if the option “all chromosomes” is chosen.
Particularly, a comparison of Holstein and Fleckvieh cattle suggested similar recombination activity genome-wide. As an example, an inspection of hotspot intervals on chromosome 3 (Figure 7) underlined regions of increased recombination rates at the chromosome ends that coincided well in both breeds. Furthermore, for each chromosome, genetic-map functions were almost overlapping. Small deviations of genetic-map functions were observed on chromosomes 5, 6, 9, 15, 16, and 18 with a slightly steeper curve in Fleckvieh.
FIGURE 7
5 Discussion and outlook
The CLARITY app provides an environment to interactively explore the physical and genetic map in selected cattle breeds. Importantly, processing of genotype data and presenting results via Shiny app rely on the current bovine genome assembly ARS-UCD1.2. In case of new assembly releases, Steps 1–3 need to be re-run and an app update becomes necessary. (Especially, both approaches in Step 2 depend on the ordering of markers for inferring phases of sire genotypes and recombination events.) Though a pipeline for Steps 1–3 is available at github, and it could theoretically be part of the Shiny app, we do not recommend its inclusion for computational matters as mentioned in Section 2.
Further work on the integration of data from other breeds (beef, dairy, dual-purpose) is underway and will facilitate complex comparative analyses of map features (e.g., hotspot intervals and assembly flaws) in different genomes. Our Shiny app will be extended accordingly, certainly increasing its value for educational and research purposes.
Statements
Data availability statement
CLARITY is a publicly available Shiny app that can be accessed via web interface at https://nmelzer.shinyapps.io/clarity. The corresponding R package CLARITY v1.0.1 including the source code and processed data can be downloaded from https://github.com/nmelzer/CLARITY under the terms of GPL (≥ 2.0). A pipeline for processing genotype data and an R script for composing the app input data are available at https://github.com/wittenburg/hsrecombi. Restrictions apply to the availability of the original data supporting the findings of this study due to thirdparty ownership. Genotype data are available from the Association for Bioeconomy Research (FBF, Bonn) and ZuchtData (Vienna) upon agreement. Requests to access the original datasets should be directed to www.fbf-forschung.de/kontakt.html; www.rinderzucht.at/zuchtdata/team.html.
Author contributions
NM developed the Shiny app package. SQ, XD, and DW analysed the data. NM, SQ, and DW wrote the manuscript and contributed to the online documentation of the R Shiny app. DW conceived the project. All authors contributed to the article and approved the final version.
Funding
This study was supported by the grant from the German Federal Ministry of Education and Research (BMBF, FKZ 031L0166 CompLS). The publication of this article was funded by the Open Access Fund of the FBN.
Acknowledgments
We gratefully acknowledge the support of the Association for Bioeconomy Research (FBF, Bonn, Germany) as representative of German Holstein cattle breeders for participating in this project and the German Evaluation Center (VIT, Verden, Germany) for composing the Holstein data. We thank ZuchtData (Vienna, Austria) for providing the German/Austrian Fleckvieh data.
Conflict of interest
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
References
1
Aden-BuieG. (2022). Metathis: HTML metadata tags for 'R markdown' and 'shiny'. Available at: https://CRAN.R-project.org/package=metathis (Accessed October 20, 2022).
2
AkyolT. Y. (2019). RVenn: Set operations for many sets. Available at: https://CRAN.R-project.org/package=RVenn (Accessed October 20, 2022).
3
AllaireJ. J. (2020). config: Manage environment specific configuration values. Available at: https://CRAN.R-project.org/package=config (Accessed October 20, 2022).
4
AttaliD. (2021). shinyjs: Easily improve the user experience of your shiny apps in seconds. Available at: https://CRAN.R-project.org/package=shinyjs (Accessed October 20, 2022).
5
AuguieB. (2017). gridExtra: Miscellaneous functions for "grid" graphics. Available at: https://CRAN.R-project.org/package=gridExtra (Accessed October 20, 2022).
6
BacheS. M.WickhamH. (2022). Magrittr: A forward-pipe operator for R. Available at: https://CRAN.R-project.org/package=magrittr (Accessed October 20, 2022).
7
ChangW.Borges RibeiroB. (2021). shinydashboard: Create dashboards with 'shiny'. Available at: https://CRAN.R-project.org/package=shinydashboard (Accessed October 20, 2022).
8
ChangW. (2021). cachem: Cache R objects with automatic pruning. Available at: https://CRAN.R-project.org/package=cachem (Accessed October 20, 2022).
9
ChangW.ChengJ.AllaireJ. J.SievertC.SchloerkeB.XieY.et al (2021). shiny: Web application framework for R. Available at: https://CRAN.R-project.org/package=shiny (Accessed October 20, 2022).
10
ChengJ.SievertC.SchloerkeB.ChangW.XieY.AllenJ. (2021). htmltools: Tools for HTML. Available at: https://CRAN.R-project.org/package=htmltools (Accessed October 20, 2022).
11
De los Ríos-PérezL.VerleihM.ReblA.BrunnerR.NguinkalJ. A.KlosaJ.et al (2020). An ultra-high density SNP-based linkage map for enhancing the pikeperch (Sander lucioperca) genome assembly to chromosome-scale. Sci. Rep.10, 22335. 10.1038/s41598-020-79358-z
12
Di FilippoL.RighelliD.GagliardiM.MatarazzoM. R.AngeliniC. (2019). HiCeekR: A novel shiny app for hi-C data analysis. Front. Genet.10, 1079. 10.3389/fgene.2019.01079
13
Everts-van der WindA.LarkinD. M.GreenC. A.ElliottJ. S.OlmsteadC. A.ChiuR.et al (2005). A high-resolution whole-genome cattle–human comparative map reveals details of mammalian chromosome evolution. PNAS102, 18526–18531. 10.1073/pnas.0509285102
14
FayC.GuyaderV.RochetteS.GirardC. (2022). Golem: A framework for robust shiny applications. Available at: https://CRAN.R-project.org/package=golem (Accessed October 20, 2022).
15
FelsensteinJ. (1979). A mathematically tractable family of genetic mapping functions with different amounts of interference. Genetics91, 769–775. 10.1093/genetics/91.4.769
16
FerdosiM. H.KinghornB. P.van der WerfJ. H.LeeS. H.GondroC. (2014). hsphase: an R package for pedigree reconstruction, detection of recombination events, phasing and imputation of half-sib family groups. BMC Bioinf15, 172. 10.1186/1471-2105-15-172
17
GaoC.-H. (2021). ggVenndiagram: A 'ggplot2' implement of Venn diagram. Available at: https://CRAN.R-project.org/package=ggVennDiagram (Accessed October 20, 2022).
18
GeorgesM.CharlierC.HayesB. (2019). Harnessing genomic information for livestock improvement. Nat. Rev. Genet.20, 135–156. 10.1038/s41576-018-0082-2
19
Google Chrome Team (2022). Squoosh!. Available at: https://github.com/GoogleChromeLabs/squoosh (Accessed June 21, 2022).
20
Google LLC (2022). Lighthouse. Available at: https://github.com/GoogleChrome/lighthouse (Accessed June 21, 2022).
21
HaldaneJ. B. S. (1919). The combination of linkage values and the calculation of distances between the loci of linked factors. J. Genet.8, 299–309.
22
HenryL.WickhamH. (2022b). purrr: Functional programming tools. Available at: https://CRAN.R-project.org/package=purrr (Accessed October 20, 2022).
23
HenryL.WickhamH. (2022a). rlang: Functions for base types and Core R and 'tidyverse' features. Available at: https://CRAN.R-project.org/package=rlang (Accessed October 20, 2022).
24
JohnssonM.JungnickelM. K. (2021). Evidence for and localization of proposed causative variants in cattle and pig genomes. Genet. Sel. Evol.53, 67. 10.1186/s12711-021-00662-x
25
LibermanU.KarlinS. (1984). Theoretical models of genetic map functions. Theor. Popul. Biol.25, 331–346. 10.1016/0040-5809(84)90013-3
26
LohP.-R.DanecekP.PalamaraP. F.FuchsbergerC.ReshefY. A.FinucaneH. K.et al (2016). Reference-based phasing using the haplotype reference consortium panel. Nat. Genet.48, 1443–1448.
27
MaL.O’ConnellJ. R.VanRadenP. M.ShenB.PadhiA.SunC.et al (2015). Cattle sex-specific recombination and genetic control from a large pedigree analysis. PLoS Genet.11, e1005387. 10.1371/journal.pgen.1005387
28
MusaA. A. (2021). “A similarity matrix and its application in genomic selection for hedging haplotype diversity (Dissertation)” (Germany: University Kiel).
29
PebesmaE. (2018). Simple features for R: Standardized support for spatial vector data. R J.10 (1), 439–446. 10.32614/RJ-2018-009
30
PurcellS.NealeB.Todd-BrownK.ThomasL.FerreiraM. A. R.BenderD.et al (2007). PLINK: a tool set for whole-genome association and population-based linkage analyses. Am. J. Hum. Genet.81, 559–575.
31
QanbariS. (2020). On the extent of linkage disequilibrium in the genome of farm animals. Front. Genet.10, 1304. 10.3389/fgene.2019.01304
32
QanbariS.SchnabelR. D.WittenburgD. (2022). Evidence of rare misassemblies in the bovine reference genome revealed by population genetic metrics. Anim. Genet.53, 498–505. 10.1111/age.13205
33
QanbariS.WittenburgD. (2020). Male recombination map of the autosomal genome in German Holstein. Genet. Sel. Evol.52, 73. 10.1186/s12711-020-00593-z
34
R Core Team (2022). R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. Available at: https://www.R-project.org.
35
RaoD. C.MortonN. E.LindstenJ.HulténM.YeeS. (1977). A mapping function for man. Hum. Hered.27, 99–104. 10.1159/000152856
36
RosenB. D.BickhartD. M.SchnabelR. D.KorenS.ElsikC. G.TsengE.et al (2020). De novo assembly of the cattle reference genome with single-molecule sequencing. GigaScience9, giaa021. 10.1093/gigascience/giaa021
37
SaliA.AttaliD. (2020). shinycssloaders: Add loading animations to a 'shiny' output while it's recalculating. Available at: https://CRAN.R-project.org/package=shinycssloaders (Accessed October 20, 2022).
38
SievertC. (2020). Interactive web-based data visualisation with R, plotly, and shiny. Boca Raton: Chapman and Hall/CRC. Available at: https://plotly-r.com.
39
WickhamH.DanenbergP.CsárdiG.EugsterM. (2021). roxygen2: In-Line Documentation for R. Available at: https://CRAN.R-project.org/package=roxygen2 (Accessed October 20, 2022).
40
WickhamH.FrançoisR.HenryL.MüllerK. (2022). dplyr: A grammar of data manipulation. Available at: https://CRAN.R-project.org/package=dplyr (Accessed October 20, 2022)
41
WickhamH. (2016). ggplot2: Elegant graphics for data analysis. New York, NY: Springer.
42
WittenburgD. (2020). hsrecombi: Estimation of recombination rate and maternal LD in half-sibs. Available at: https://cran.r-project.org/package=hsrecombi (Accessed October 20, 2022).
43
WomackJ. E. (2005). Advances in livestock genomics: Opening the barn door. Genome Res.15, 1699–1705. 10.1101/gr.3809105
44
XieY.ChengJ.TanX. (2022). DT: A wrapper of the JavaScript library 'DataTables'. Available at: https://CRAN.R-project.org/package=DT (Accessed October 20, 2022).
Summary
Keywords
single nucleotide polymorphism, linkage, recombination rate, education, mapping function
Citation
Melzer N, Qanbari S, Ding X and Wittenburg D (2023) CLARITY: a Shiny app for interactive visualisation of the bovine physical-genetic map. Front. Genet. 14:1082782. doi: 10.3389/fgene.2023.1082782
Received
28 October 2022
Accepted
12 May 2023
Published
30 May 2023
Volume
14 - 2023
Edited by
Joanna Szyda, Wroclaw University of Environmental and Life Sciences, Poland
Reviewed by
Arvind Sonwane, Indian Council of Agricultural Research (ICAR), India
Nathan Olson, National Institute of Standards and Technology (NIST), United States
Updates
Copyright
© 2023 Melzer, Qanbari, Ding and Wittenburg.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Dörte Wittenburg, wittenburg@fbn-dummerstorf.de
Disclaimer
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.