Carhta Gene: multipopulation integrated genetic and radiation hybrid mapping (original) (raw)

Journal Article

,

INRA, Biométrie et Intelligence Artificielle/Génétique Cellulaire BP 27, 31326 Castanet-Tolosan Cedex, France

*To whom correspondence should be addressed.

Search for other works by this author on:

,

INRA, Biométrie et Intelligence Artificielle/Génétique Cellulaire BP 27, 31326 Castanet-Tolosan Cedex, France

Search for other works by this author on:

,

INRA, Biométrie et Intelligence Artificielle/Génétique Cellulaire BP 27, 31326 Castanet-Tolosan Cedex, France

Search for other works by this author on:

,

INRA, Biométrie et Intelligence Artificielle/Génétique Cellulaire BP 27, 31326 Castanet-Tolosan Cedex, France

Search for other works by this author on:

INRA, Biométrie et Intelligence Artificielle/Génétique Cellulaire BP 27, 31326 Castanet-Tolosan Cedex, France

Search for other works by this author on:

Received:

14 September 2004

Revision received:

03 November 2004

Accepted:

03 November 2004

Published:

14 December 2004

• PDF
• Split View
• • Article contents
  • Figures & tables
  • Video
  • Audio
  • Supplementary Data
• Cite

Cite

Simon de Givry, Martin Bouchez, Patrick Chabrier, Denis Milan, Thomas Schiex, Carhta Gene: multipopulation integrated genetic and radiation hybrid mapping, Bioinformatics, Volume 21, Issue 8, April 2005, Pages 1703–1704, https://doi.org/10.1093/bioinformatics/bti222
Close

Navbar Search Filter Mobile Enter search term Search

Abstract

Summary: Carhta Gene: is an integrated genetic and radiation hybrid (RH) mapping tool which can deal with multiple populations, including mixtures of genetic and RH data. Carhta Gene: performs multipoint maximum likelihood estimations with accelerated expectation–maximization algorithms for some pedigrees and has sophisticated algorithms for marker ordering. Dedicated heuristics for framework mapping are also included. Carhta Gene: can be used as a C++ library, through a shell command and a graphical interface. The XML output for companion tools is integrated.

Availability: The program is available free of charge from www.inra.fr/bia/T/CarthaGene for Linux, Windows and Solaris machines (with Open Source).

Contact: tschiex@toulouse.inra.fr

INTRODUCTION

The genetic mapping technique is used to locate polymorphic markers on chromosomes by making use of a probabilistic model of crossing-over. Many genetic mapping tools are available to analyze data in experimental crosses. Most of them are designed to analyze line crosses, one family at a time; few can integrate data from several crosses to build consensus maps (see http://linkage.rockfeller.edu/soft).

Radiation hybrid (RH) mapping is a somatic cell technique that complements the genetic mapping technique, allowing for finer resolution. Most existing RH mapping packages are listed at compgen.rutgers.edu/rhmap

Although capable of handling line crosses, Carhta Gene has been designed to create consensus maps from multiple populations. Instead of directly integrating existing maps, Carhta Gene computes maximum multipoint likelihood maps, taking into account all the available information, offering additional reliability. It can integrate RH and genetic data together.

METHODS

Parametric probabilistic HMM based models for crossover during meiosis (Lander et al., 1987) and for chromosome breakage and retention during RH panel construction (Lange et al., 1995) include parameters such as recombination and breakage probabilities between adjacent markers as well as retention probability. Given experimental data, and assuming some marker ordering, the values of these parameters can be estimated by maximizing the probability that the data observed has been generated by the model. This likelihood therefore allows simultaneous evaluation of the assumed marker ordering and estimation of the corresponding parameters (distances). This is done by a so-called expectation–maximization (EM) algorithm. Models for backcross, f2 intercross, recombinant inbred lines (self and sibs) and phase-known outbreds are available. For RH estimation, the equal retention model both in its haploid and diploid forms is used. The EM forward–backward algorithm used in Carhta Gene has been accelerated by taking into account specific properties of backcross and haploid RH data (see Schiex et al., 2001). Compared with usual EM implementations, the accelerated algorithm can run one or two orders of magnitude faster with no loss of precision.

Carhta Gene provides two ways to merge data files:

1. If one assumes that the data merged represents a single map (same order and distances, either genetic or RH), a so-called genetic merging is done. Untyped markers in a population/panel are considered as missing data and one consensus map is produced. RH and genetic data cannot be merged under this model because they use different distances.
2. Otherwise, it is assumed that the data files are representative of maps with a common marker ordering but specific distances per dataset. Here, a single consensus ordering is produced but a specific set of distances is estimated for each model merged. Any type of data, genetic or RH, can be merged here. This is called order merging. These two methods can be combined freely.

For genetic merging, the EM implementation deals with combined datasets by performing an E computation on each dataset and then using an M step that takes into account the merging performed. For datasets merged by order, independent log-likelihoods are obtained per dataset and summed up.

The main problem in genetic or RH mapping arises from the number of possible marker orders. For n markers, there exists n!/2 different possible orders. Since the connection between mapping and the traveling salesman problem (TSP) is well known for genetic mapping (Schiex and Gaspin, 1997www.inra.fr/bia/T/CartaGene) and for RH mapping, (Ben-Dor et al., 2000) Carhta Gene relies on this connection and provides extensions of TSP solving algorithms:

• Fast heuristics (e.g. nearest neighbor) using 2-point information for guidance and multipoint estimations for final evaluation.
• More powerful meta-heuristics (e.g. simulated annealing and taboo search) directly using multipoint maximum likelihood as the optimization criterion.
• Even more efficient pure TSP heuristics such as the Lin-Kernighan heuristic (LKH) (Lin and Kernighan, 1973), using the LKH implementation (Helsgaun, 2000, http://www.dat.ruc.dk/keld/research/LKH). These algorithms can use either multiple 2-point maximum likelihood or obligate chromosome breaks (Ben-Dor et al., 2000; Agarwala et al., 2000) to define optimal maps. All the maps identified in this way are then evaluated using a multipoint EM based estimation.
• Finally, dedicated heuristics for framework mapping, duplicated marker detection and map validation.

A unique feature of Carhta Gene is that, instead of producing a single supposedly optimal map, it produces an ordered set of alternative maps which allows an estimation of the reliability of ordering of each marker. The set of all these maps can be explored manually and compared graphically (with a Postscript output; Fig. 1). Dedicated automatic tools facilitate the identification of unreliable markers for further analysis.

Final maps can be produced under MapMaker (Lander et al., 1987) and XML formats for data exchange with MCQTL (Jourjon et al., 2004) (a multiple population QTL mapping software) and BioMercator (Arcade et al., 2004) (a software for integrating genetic maps and QTL detected in independent experiments).

IMPLEMENTATION

Carhta Gene is implemented as a C++ library. A Tcl programable shell command for automated mapping is available and a graphical Tcl/Tk interface for interactive mapping. Binaries for Windows, Solaris and Linux are provided with an open source distribution (using the G-Forge site mulcyber.toulouse.inra.fr).

Fig. 1

A typical graphical session of CarHTa Gene:, displaying possible maps with distance and log-likelihoods.

This work was supported by GENOPLANTE project ‘Integrative Tools for Genetic Mapping’.

REFERENCES

Agarwala, R., Applegate, D.L., Maglott, D., Schuler, G.D., Schaffer, A.A.

2000

A fast and scalable radiation hybrid map construction and integration strategy.

Genome Res.

10

350

– 364

Arcade, A., Labourdette, A., Falque, M., Mangin, B., Chardon, F., Charcosset, A., Joets, J.

2004

Biomercator: integrating genetic maps and QTL towards discovery of candidate genes.

Bioinformatics

20

2324

– 2326

Ben-Dor, A., Chor, B., Pelleg, D.

2000

RHO—radiation hybrid ordering.

Genome Res.

10

365

– 378

Helsgaun, K.

2000

An effective implementation of the Lin-Kernighan traveling salesman heuristic.

Eur. J. Oper. Res.

126

106

– 130

Jourjon, M.-F., Jasson, S., Marcel, J., Ngom, B., Mangin, B.

2004

MCQTL: Multi-allelic QTL mapping in multi-cross design.

Bioinformatics

21

128

– 130

Lander, E.S., Green, P., Abrahamson, J., Barlow, A., Daly, M.J., Lincoln, S.E., Newburg, L.

1987

MAPMAKER: an interactive computer package for constructing primary genetic linkage maps of experimental and natural populations.

Genomics

1

174

– 181

Lange, K., Boehnke, M., Cox, D.R., Lunetta, K.L.

1995

Statistical methods for polyploid radiation hybrid mapping.

Genome Res.

5

136

– 150

Lin, S. and Kernighan, B.W.

1973

An effective heuristic algorithm for the traveling salesman problem.

Oper. Res.

21

498

– 516

Schiex, T. and Gaspin, C.

1997

Cartagene: constructing and joining maximum likelihood genetic maps.

Proceedings of ISMB'97

, Halkidiki, Greece Porto Carras, pp.

258

–267

Schiex, T., Chabrier, P., Bouchez, M., Milan, D.

2001

Boosting EM for radiation hybrid and genetic mapping.

Proceedings of WABI'01

, pp.

41

–51 vol. 2149 LNCS

© The Author 2004. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oupjournals.org

Citations

Views

Altmetric

Metrics

Total Views 1,759

1,350 Pageviews

409 PDF Downloads

Since 12/1/2016

Citations

314 Web of Science

×

Email alerts

Citing articles via

• Latest

• Most Read

• Most Cited

More from Oxford Academic