A Consensus Map for Loblolly Pine (Pinus taeda L.). I. Construction and Integration of Individual Linkage Maps From Two Outbred Three-Generation Pedigrees

Corresponding author: David B. Neale, Institute of Forest Genetics, USDA Forest Service, 1 Shields Ave., Department of Environmental Horticulture, University of California, Davis, CA 95616., dneale{at}dendrome.ucdavis.edu (E-mail)

Communicating editor: A. H. D. BROWN

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

A consensus map for loblolly pine (Pinus taeda L.) was constructed from the integration of linkage data from two unrelated three-generation outbred pedigrees. The progeny segregation data from restriction fragment length polymorphism, random amplified polymorphic DNA, and isozyme genetic markers from each pedigree were recoded to reflect the two independent populations of parental meioses, and genetic maps were constructed to represent each parent. The rate of meiotic recombination was significantly greater for males than females, as was the average estimate of genome length for males {1983.7 cM [Kosambi mapping function (K)]} and females [1339.5 cM(K)]. The integration of individual maps allows for the synthesis of genetic information from independent sources onto a single consensus map and facilitates the consolidation of linkage groups to represent the chromosomes (n = 12) of loblolly pine. The resulting consensus map consists of 357 unique molecular markers and covers ~1300 cM(K).

A genetic linkage map is a fundamental organizational tool for genomic research. For forest trees, the most important applications of genetic maps are toward (1) a basic knowledge of genome organization and evolution (KINLAW and NEALE 1997 ), (2) the localization of monogenic traits (e.g., DEVEY et al. 1995 ) and oligogenic traits (e.g., GROOVER et al. 1994 ), and (3) studies of genetic diversity (MITTON 1994 ). Therefore, for any given species, individual genetic maps are often constructed with a specific goal in mind, thereby generating multiple maps for a single species that feature novel markers and genetic information. The information contained within these individual maps can be further enhanced when these maps are synthesized into a single consensus map to represent a given species.

Consensus maps have been constructed for a number of plant species (e.g., Arabidopsis thaliana, HAUGE et al. 1993 ; Brassica oleracea, KIANIAN and QUIROS 1992 ; Helianthus annus, GENTZBITTEL et al. 1995 ; Hordeum vulgare, QI et al. 1996 ; Zea mays, BEAVIS and GRANT 1991 ). Mapping with multiple populations provides several advantages over mapping based on a single population. In particular, a larger number of loci can be placed onto a single map. This is especially important when attempting to map specific genes of interest (e.g., morphological markers or candidate genes for trait mapping) that are unlikely to segregate within a single mapping population. This also provides for greater genomic coverage. These multipopulation mapping studies have provided evidence for chromosomal rearrangements (BEAVIS and GRANT 1991 ; KIANIAN and QUIROS 1992 ) and gene duplication (KIANIAN and QUIROS 1992 ; GENTZBITTEL et al. 1995 ), have assisted in the assignment of linkage groups to chromosomes (BEAVIS and GRANT 1991 ), and have provided the basis for comparative studies among related species and subspecies (KIANIAN and QUIROS 1992 ; HAUGE et al. 1993 ; GENTZBITTEL et al. 1995 ).

Loblolly pine (Pinus taeda L.) is a member of the genus Pinus, which encompasses ~100 species and is among the world's most widely distributed and economically important group of trees. Pines are characterized by their longevity, outbred mating system, and high genetic variability relative to other plant species (HAMRICK and GODT 1990 ). Because of its wide range, rapid growth, and versatility as a source for wood products, loblolly pine is the principal commercial pine species in North America. All members of the genus Pinus have a haploid chromosome number of 12. Compared to other plant species, pine genomes are large (e.g., the estimated C-value for loblolly pine is 21–23 pg; WAKAMIYA et al. 1993 ) and are composed of complex gene families (KINLAW and NEALE 1997 ). Nevertheless, initial investigations suggest that gene order has been conserved among pines (DEVEY et al. 1996B ), and cytogenetic studies conclude that Pinus has been cytologically very stable over evolutionary time (PEDERICK 1970 ).

For the majority of plant species of academic and agronomic interest, pedigrees involving backcrosses (BCs) or F₂s from inbred lines are constructed for mapping purposes. This pedigree structure simplifies the mapping procedures because the genetic segregation observed in the mapping population is the result of the meiotic recombination from a single genotype (the F₁). Therefore, only two alleles segregate in these mapping populations, and the resulting genetic map represents the recombination of the F₁ parent. Furthermore, because two homozygous inbred lines are used to generate the F₁ parent, linkage phase among the alleles of the mapping population is known a priori.

In contrast, high genetic load typically prevents the construction of inbred lines in coniferous forest tree species (FRANKLIN 1970 ). Consequently, both parents of an outbred pedigree are typically highly heterozygous and can possess different pairs of alleles at each locus (i.e., as many as four alleles can segregate for any given marker). Therefore, the genetic segregation observed in such mapping populations is the result of meiotic recombination from both parents, and any given marker can segregate in two (1:1), three (1:2:1), or four (1:1:1:1) genotypic classes within a single mapping population. In addition, phase relationships among alleles are not known a priori, but instead must be determined either from the inheritance of alleles within a three-generation pedigree structure or from progeny segregation data.

With a cross between two highly heterozygous parents, the progeny data can be subdivided into two independent data sets that separately contain the meiotic segregation data from each parent. Independent linkage maps can then be constructed for each parent. This "single-parent" approach is an extension of the strategy used to construct maternal linkage maps from random amplified polymorphic DNA (RAPD) markers and haploid megagametophytes in conifers (TULSIERAM et al. 1992 ). Furthermore, the codominant and multiallelic nature of restriction fragment length polymorphism (RFLP) markers provides orthologous markers for aligning the linkage information from each parental data set to produce a sex-average map (DEVEY et al. 1994 ).

As an initial step toward synthesizing the genetic information available for loblolly pine, the present study integrates the linkage data from two independent outbred pedigrees into a single consensus map. This consensus map serves as a reference genetic map for loblolly pine and as a foundation for studies of genome organization and evolution in conifers.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Data collection
Mapping populations: Two three-generation outbred pedigrees of loblolly pine were constructed from unrelated first-generation selections of the North Carolina State University-Industry Cooperative Tree Improvement Program. These pedigrees were constructed and maintained by the Weyerhauser Company. The first pedigree (referred to as the base pedigree) consists of 95 full-sib progeny and was previously used to construct a genetic map from 90 RFLP and 6 isozyme loci (DEVEY et al. 1994 ). The second pedigree (referred to as the qtl pedigree) consists of 172 full-sib progeny. GROOVER et al. 1994 previously used a selective genotyping strategy to construct the initial genetic map for the qtl pedigree. Therefore, a subset of 48 extreme-phenotype progeny was used as the preliminary mapping population. Genotypic data for the 124 intermediate-phenotype progeny were subsequently obtained for selected markers throughout the genome. A total of 155 RFLP loci were used for this previous map construction (GROOVER et al. 1994 ).

Genetic markers: Three sources of genetic markers (RFLPs, RAPDs, and isozymes) were used for each mapping population. Methods pertaining to RFLP analysis for loblolly pine were described in DEVEY et al. 1991 . The primary source of RFLP probes was a loblolly pine complementary DNA (cDNA) library constructed by DEVEY et al. 1991 . Probes from a genomic DNA library were also used (DEVEY et al. 1991 ). In addition, various colleagues have contributed a number of gene probes characterized from loblolly pine and P. sylvestris. The RAPD analysis followed methods described in NELSON et al. 1993 . Only RAPD markers in a BC configuration for either parent were used in linkage analyses. Methods for protein extraction from vegetative bud tissue for isozyme analyses were described in NEALE et al. 1984 . Gel electrophoresis and protein staining followed the method of CONKLE et al. 1982 . The protein stain for cinnamyl alcohol dehydrogenase followed the method of O'MALLEY et al. 1992 .

Data organization: The original base and qtl maps were constructed independently, and little effort was made to assign common names to redundant probes or to identify orthologous markers. The positive identification of orthologous RFLP markers is not straightforward in conifers because of the presence of multiple alleles and multigene families, yet it is fundamental to the effective utilization and synthesis of genetic data from independent sources. Therefore, to organize and integrate these independent data sets, the raw segregation data were completely rescored, and band migration distances were visually compared as a means to identify orthologous markers and alleles within and among each pedigree data set. This identification procedure was twofold: (1) to identify sets of redundant probes (the final data set includes only a single representative from each set of redundant probes); and then (2) among members of a multigene family (revealed either by nonredundant or redundant probes), to identify putative orthologous markers among individual mapping populations. Once these data sets were organized and the orthologous markers were identified, the probes that were analyzed in only one pedigree were applied to the other pedigree as a means to identify additional orthologous markers.

Segregation and linkage analyses
Genotypic classification and coding: Raw genotypic data were initially recorded with no previous knowledge of phase relationships or inheritance. By using the three-generation pedigree structure, the raw data were converted into an "ancestry-known" data format that resembles that of an inbred pedigree. First, the alleles for each progeny and parent genotype were reoriented to reflect their inheritance from the previous generation. Second, the data were divided into parental lines, and two gametic segregation data sets were created to represent the populations of segregating meioses from each parent. In situations where the ancestry could not be determined (i.e., when a pair of grandparents were heterozygous for the same pair of alleles or when grandparental data were missing), the data remained in an "ancestry-unknown" format.

Data sets were then coded for linkage analyses with MAPMAKER (Macintosh version 2.0; LANDER et al. 1987 ), MAPMANAGER (Macintosh version 2.6.1; MANLY 1993 ), and JOINMAP (UNIX version 1.4; STAM 1993 ). When marker data from an outbred pedigree are oriented to reflect ancestry for each parental line, the coding designations (i.e., A and H) for progeny segregation data represent alleles that were inherited from the grandmother or grandfather, respectively, rather than "homozygote" or "heterozygote." Therefore, the ancestry-known markers can be analyzed within a phase-known model. To accommodate the remaining ancestry-unknown data within a phase-known model, these markers were "reciprocally" coded to arbitrarily represent both possible ancestry options (NELSON et al. 1993 ). During the linkage analyses, the correct phase of the ancestry-unknown markers was established relative to the ancestry-known markers. The original data set was then updated to include the correct phase coding for the ancestry-unknown markers, producing a complete data set that is both ancestry and phase known.

Because JOINMAP manages phase-unknown data internally, the reciprocal coding step described above is not necessary. The disadvantage of this internal manipulation is that the phase orientation that is eventually used is not conveniently made known to the user. In addition, because MAPMAKER and MAPMANAGER cannot simultaneously analyze markers in both a BC and intercross (IC) configuration without a loss of data points, the markers in an IC configuration were removed from each data set for MAPMAKER and MAPMANAGER analyses.

Concurrent linkage analyses of multiple populations: Chi-square values were calculated for individual markers to detect deviation (P 0.05) of gametic segregation from Mendelian ratios (1:1). Linkage analyses for each mapping population were initially performed using MAPMAKER. Linked markers were first placed into linkage groups using the "group" command with LOD 5.0 and recombination fraction () 0.30. A preliminary estimation of marker order was obtained using the "first order" command. These preliminary linkage groups from each mapping population were then visually aligned via orthologous markers and coalesced into sets of linkage group homologues. Linkage data within individual mapping populations were consequently joined into individual linkage groups by LOD 5.0 and 0.30 and/or orthologous markers. Additional markers were then placed onto these linkage groups at LOD 3.0 and 0.40. This strategy ensures confidence in the formation of each set of linkage group homologues, yet it allows the placement of a maximum number of markers as well as the proper orientation of linkage fragments. The probable order of all markers for each linkage group was determined using the "compare" command and verified using the "ripple" command. The marker segregation data for these orders were visually inspected using the program MAPMANAGER. The IC markers were then placed relative to the established order of BC markers using JOINMAP and the "fixed sequence" command.

Integration of linkage data: An integrated linkage analysis was performed using all segregation data (i.e., both BC and IC markers) from the four mapping populations using JOINMAP. Pairwise estimates of and corresponding LOD scores were calculated from segregation data for each population. When orthologous markers defined an interval for two or more populations, JOINMAP replaced the individual values of with a weighted average value (STAM 1993 ). Using the fixed sequence command, these data were integrated relative to the established order of BC markers ascertained from the individual MAPMAKER analyses. Map units in centimorgans were derived from the Kosambi (K) mapping function and were used to construct all linkage maps from both MAPMAKER (results not shown) and JOINMAP analyses. Map figures were constructed using the computer program Linkage Group Visualization Tool (P. SKAGGS, unpublished results).

Comparison of meiotic recombination rate among parents: Orthologous marker pairs were chosen for sequential intervals spaced evenly along linkage groups. A minimum of 40 progeny were used in pairwise analyses. A test for a global difference in maternal and paternal recombination frequency among parental data sets was performed from averages of marker pair intervals from each pedigree using a z-test with the delta method to estimate the associated standard errors (BISHOP et al. 1975 ).

Estimation of genome length and map coverage: ECHT and NELSON 1997 identified the need for adopting uniform criteria when estimating genome length. Following these criteria, each data set was reduced to a progeny size of 48 to avoid missing data points but to still include a random sample of all markers. Estimations of genome length, E(G) (HULBERT et al. 1988 ), and expected genome coverage, E(C_n)% (BISHOP et al. 1983 ), were calculated from pairwise segregation data for marker pairs above threshold LODs of 3, 4, and 5. Confidence intervals for genome length were calculated following the method of GERBER and RODOLPHE 1994 . The observed genome coverage, C_obs, was calculated from the results of map construction using a 30-cM scale (NELSON et al. 1994 ). These procedures are detailed in ECHT and NELSON 1997 .

Nomenclature and database: Locus nomenclature followed that outlined on the Genome Resources page in Dendrome, the forest tree genome informatics project website (http://dendrome.ucdavis.edu/). A modification of the alphanumeric coding was used for the locus identifier field for RFLP markers. A lowercase letter was used for RFLP markers in the base pedigree, whereas a number was used for RFLP markers in the qtl pedigree. For RFLP markers found in both pedigrees, an uppercase letter was used to indicate commonality and facilitate map integration. An additional prefix may be added to a locus or mapping population name, e.g., LG2/2986.A or LG4/qtl-pat, respectively, to indicate a specific linkage group.

The constituent and consensus maps (SEWELL and NEALE 1998 ) and all information pertaining to map construction are located in the Dendrome and/or TreeGenes databases (http://dendrome.ucdavis.edu/Treegenes/). The cDNA clones used for RFLP analyses are freely distributed and may be obtained by submitting a request to dneale@dendrome.ucdavis.edu via electronic mail.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Genetic markers and single-locus segregation analysis:
From 168 RFLP probes, 42 RAPD primers, and 12 isozyme stains, a total of 581 genetic markers that segregated among the two independent outbred pedigrees of loblolly pine were identified. These marker data were subdivided into four data sets: base-maternal (base-mat), base-paternal (base-pat), qtl-maternal (qtl-mat), and qtl-paternal (qtl-pat). Of these 581 markers (430 cDNA RFLPs, 36 genomic RFLPs, 87 RAPDs, and 28 isozymes), a total of 418 unique markers (300 cDNA RFLPs, 23 genomic RFLPs, 80 RAPDs, and 15 isozymes) were identified. The remaining 163 markers were orthologous complements found in two or more of the constituent data sets. On average among these two pedigrees, each marker type yielded 1.88 unique segregating loci (1.89 per cDNA probe, 2.55 per genomic probe, 1.90 per RAPD primer, and 1.25 per isozyme stain). The maximum number of scorable segregating loci per marker type was eight (cDNA probe), six (genomic probe), three (RAPD primer), and two (isozyme stain).

Allelic associations among RFLP fragments could not be identified for some loci (Figure 1). Consequently, these loci were scored as presence vs. absence for allele fragments. This phenomenon has been observed in other plant species and in other conifers (e.g., Arabidopsis thaliana, LIU et al. 1996 ; B. oleracea, KIANIAN and QUIROS 1992 ; H. annus, GENTZBITTEL et al. 1995 ; P. radiata, DEVEY et al. 1996A ; Pseudotsuga menziesii, JERMSTAD et al. 1994 ). Several factors could account for this. The "null" allele might go undetected as a result of comigration with another fragment or limitations of the technique of analysis. Alternatively, the locus might possess a null allele and/or be regarded as hemizygous.

View larger version (69K): In this window In a new window Download PPT slide   Figure 1. Autoradiogram from a Southern blot analysis of the qtl pedigree. Genomic DNAs from 23 progeny were hybridized to cDNA probe PtIFG 2899. RFLP fragments 41, E1, 61, and 71 were scored as present vs. absent, and they are treated as dominant markers. Loci 2899 A, 2899 4, 2899 E, and 2899 7 are dispersed throughout the genome (LG12, LG1, LG3, and LG5, respectively; see Figure 2). Locus 2899 6 is in an IC configuration and was not used in linkage analyses. 2899 2 and 2899 3 are putative loci that did not segregate in either the base or qtl pedigrees.

Eight markers (1.4%, two from each mapping population) exhibited significant distortion (P 0.05) from expected Mendelian ratios based on chi-square tests (results not shown). This observed single-locus segregation distortion does not appear to result from selection for or against the alleles of a particular grandparent and is likely due to chance alone.

Construction of individual linkage maps and multilocus segregation analyses:
The four data sets—base-mat, base-pat, qtl-mat, and qtl-pat—contained 109, 106, 191, and 175 genetic markers, respectively, of which 97 (89.0%), 79 (74.5%), 179 (93.7%), and 153 (87.4%) were linked at LOD 3.0 (Table 1). These 508 (87.5%) mapped markers consisted of 415 RFLPs, 73 RAPDs, and 20 isozymes. Among these markers, 108 were common to two or more populations (Table 2; SEWELL and NEALE 1998 ) and, thus, were used to identify and align linkage group homologues among mapping populations.

The maps generated by JOINMAP, which includes all markers from both BC and IC configurations, cover 699, 591, 908, and 956 cM (K) for each individual mapping population (base-mat, base-pat, qtl-mat, and qtl-pat, respectively; Table 1). These maps (see SEWELL and NEALE 1998 ) were constructed from the same marker order as defined for the BC markers by the MAPMAKER analyses (maps not shown). The maps constructed from JOINMAP are slightly shorter than those from MAPMAKER (Table 1) even though the JOINMAP analyses included the additional IC markers. This difference was also observed in barley and is attributed to how each program calculates map distance when the actual interference differs from that assumed (QI et al. 1996 ).

Orthologous markers among constituent linkage group homologues were compared for collinearity of marker order. With minor exceptions, complete collinearity was observed among the orthologous markers of these linkage groups. The minor exceptions included LG5a/base-pat, for which markers mapped to the same general area as those from other linkage group homologues, although the marker order is slightly different (SEWELL and NEALE 1998 ). This region contained a relatively high frequency of double crossovers that made conclusive ordering of markers difficult. Also, several IC markers, which are less precisely ordered because of the presence of ambiguous heterozygous progeny classes, appear inverted among intrapedigree linkage groups (e.g., LG2/base-2885 B and LG9/qtl-Got-4). These comparisons among individual maps from within a species are consistent with those among a self and F₁ map from an individual tree (PLOMION et al. 1995 ), suggesting that the reported discrepancies are caused by minor statistical errors and are not biological in nature. Therefore, we found no significant evidence of minor intraspecific chromosomal rearrangement among the four loblolly pine parent trees.

The availability of orthologous markers also allowed for a direct comparison of the rate of meiotic recombination among the four mapping populations. A comparison of mean recombination frequencies among intervals from throughout the genome revealed a significantly greater male than female meiotic recombination rate in both the base (26.12%) and qtl (26.33%) pedigrees (Table 3). This is consistent with that reported by GROOVER et al. 1995 from a smaller subset of marker pairs for these same pedigrees.

View this table: In this window In a new window   Table 3. Global test for deviation between maternal and paternal recombination rates ()

Estimates of genome length were calculated from pairwise segregation data for each mapping population (Table 4). The average estimates of genome length in cM(K) and their 95% confidence intervals (in parentheses) at LOD threshold 4 were as follows: 1315.5 (1117.9–1612.7), base-mat; 2129.1 (1696.4–2845.7), base-pat; 1363.4 (1241.1–1526.6), qtl-mat; and 1838.2 (1608.3–2075.3), qtl-pat. The average estimates for each maternal genome length were nearly identical, and those for the paternal genomes were not significantly different. However, the paternal estimates were significantly greater than the maternal estimates within each pedigree (38.2% difference within base and 25.8% within qtl) and also among pedigrees for qtl-mat and base-pat (and nearly so between base-mat and qtl-pat).

Construction of integrated consensus map:
The linkage data from these four independent mapping populations contained 108 orthologous markers that were common to at least two mapping populations. The base-mat, base-pat, qtl-mat, and qtl-pat data sets contained 55, 57, 81, and 65 orthologous markers, respectively (Table 2). The resulting consensus map (Figure 2) contains 357 (278 RFLPs, 67 RAPDs, and 12 isozymes) unique genetic markers (Table 1). Consensus linkage groups 1–12 were integrated from each of the four independent mapping populations and cover 1175 cM(K) of the genome. Consensus groups 13 and 14 were integrated only within the qtl pedigree and account for an additional 52 cM(K). The remaining six small "single-population" linkage groups cover 133 cM(K). These 20 linkage groups potentially provide a total genome coverage of 1359 cM(K). Without the presence of orthologous markers on these single-population linkage groups, however, it is undeterminable whether this 133 cM(K) is already covered by the integrated regions of the consensus map. Thus, a conservative estimate of the total genome coverage for the consensus map from unique linkage segments is 1227 cM(K). With a few minor exceptions (e.g., LG7/1636.1; SEWELL and NEALE 1998 ), the marker order on the consensus map was not different from that defined by each constituent map.

View larger version (70K): In this window In a new window Download PPT slide   Figure 2. A consensus map for loblolly pine. The constituent maps from the base and qtl pedigrees (SEWELL and NEALE 1998 ) are located on the World Wide Web at http://dendrome.ucdavis.edu/Data/.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The primary goal for the construction of this consensus map was to place, relative to one another, as many genetic markers as possible onto a single map. Therefore, the concern is more towards obtaining a general order and distance among these markers rather than the fine resolution of order and distance. Furthermore, the consensus map provides a means to consolidate linkage groups from the base and qtl pedigrees (Figure 2; SEWELL and NEALE 1998 ). With 357 unique genetic markers from cDNA and genomic probes, RAPDs, and isozyme markers, this map is an excellent resource from which markers may be selected for future mapping projects within loblolly pine and for comparative studies among other pines (e.g., DEVEY et al. 1996B ). For example, combining information from multiple pedigrees is necessary if important traits do not segregate within a single population. Often it is not practical or necessary to construct complete genetic maps to identify the genomic locations of these traits (e.g., DEVEY et al. 1995 ). However, such studies can still be related to the entire genome by selecting markers suitable for superimposing the detailed region onto the consensus map. In a similar manner, the consensus map can also serve as a reference for considering the pines as a single genetic system, as similarly proposed for the grasses (BENNETZEN and FREELING 1993 ), in which genetic tools and information from one pine species immediately become available for studying other pines.

A significant difference between female and male meiotic recombination frequencies was detected among map intervals defined by linked orthologous markers (Table 3). Although reciprocal crosses were not investigated, these results support those from P. radiata for which reciprocal crosses were analyzed (MORAN et al. 1983 ). The higher rate of male recombination was observed throughout the genome and resulted in the fragmentation of paternal linkage groups (e.g., LG1/base-pat and LG1/qtl-pat; SEWELL and NEALE 1998 ). This difference also resulted in a lower percentage of mapped markers within the paternal mapping populations relative to the maternal populations, where a similar number of markers was used (Table 1). BEAVIS and GRANT 1991 raised the question of whether data should be pooled for construction of a composite map if unequal recombination is detected among constituent populations. However, they pointed out that such a composite map is still a valuable reference tool for selecting markers and comparing the relative location of traits in different genetic backgrounds.

This difference among female and male recombination frequencies directly influenced the estimates of genome length calculated for loblolly pine because these estimates are, in part, based on the number of linked marker pairs relative to the total number of marker pairs within a mapping population. The estimated genome length for loblolly pine was 1315 and 1363 cM(K) when using maternal segregation data and 1838 and 2129 cM(K) when using paternal data (Table 4). ECHT and NELSON 1997 used the same criteria to estimate the genome length of P. palustris, P. pinaster, and P. strobus from maternal segregation data using RAPD markers. The reported genome length for each of these species, which represent both subgenera (Pinus and Strobus) of the genus Pinus, is ~2000 cM(K). Based on comparisons of the reported confidence intervals, the maternal estimate for genome length of loblolly pine is significantly different from the maternal estimates reported by ECHT and NELSON 1997 . Loblolly pine is closely related to longleaf pine; therefore, taxonomic relatedness does not appear to be a factor regarding the difference among these estimates. A possible factor that might contribute to this difference is the type of marker used for each of these estimates. Putatively, RAPD markers are more randomly dispersed throughout the genome than RFLP markers (detected from cDNA probes) and isozymes. If clustering does exist among the majority of markers used in the loblolly pine study, then genome length would be underestimated.

Nonuniform marker distributions are observed as gaps among many of the linkage groups (Figure 2; SEWELL and NEALE 1998 ). These gaps may represent a nonrandom sampling of the genome resulting from an underrepresentation of markers from these regions. Alternatively, the markers may be random along the physical length of the chromosome, and these gaps may represent an uneven distribution of recombination events along the chromosome (TANKSLEY et al. 1992 ). It is possible that the distribution of expressed genetic probes (i.e., cDNAs and isozymes) is not representative of the entire genome. RAPD markers were included in the analysis in an attempt to add markers from throughout the genome. However, the addition of RAPD markers did not bridge any gaps. This evidence suggests that the genome has been randomly sampled and the gaps represent nonuniform recombination events. Furthermore, the observation that gaps can be found in the same general region in each of the four independent maps (e.g., linkage groups 1 and 5; SEWELL and NEALE 1998 ) suggests that these gaps represent some sort of cytogenetic phenomenon (e.g., the presence of centromeres; TANKSLEY et al. 1992 ) rather than a simple genome sampling error. Notwithstanding, ~13% of the markers remain unlinked among these mapping populations.

Southern blot analyses using cDNA and gene probes have revealed genes that are found in duplicate, and occasionally multiple, copies in many plant species (e.g., BERNATZKY and TANKSLEY 1986 ; HELENTJARIS et al. 1988 ). However, the majority of cDNA and gene probes in loblolly pine detect genes in multigene families (e.g., Figure 1; KINLAW and NEALE 1997 ). These gene families are often significantly larger than their counterparts in angiosperms (KINLAW et al. 1994 ; PERRY and FURNIER 1996 ) and they are found both dispersed throughout the genome (e.g., Figure 1) or as tightly linked clusters (e.g., LG10/1635; Figure 2).

Polyploidization has played a role in the evolution of gene families in plant genomes. For example, the genetic map of maize revealed many syntenic regions that are duplicated throughout the genome, which are believed to be the result of an ancient polyploidization event, followed by genomic rearrangements (HELENTJARIS et al. 1988 ). Cytogenetic studies of pines have generated conflicting conclusions regarding polyploidization; MEHRA and KHOSHOO 1956 and MIROV 1967 find no evidence for polyploidization, whereas DREWRY 1988 supports this possibility. Currently, the consensus map provides no evidence for the presence of duplicated syntenic regions. Furthermore, gene family sizes in pines vary considerably among different species and probes (AHUJA et al. 1994 ), suggesting that gene amplification is an ongoing process in pines.

Alternatively, one of the possible mechanisms of gene amplification in conifers is reverse transcription of mRNA templates resulting in processed retropseudogenes (DROUIN and DOVER 1987 ). KAMM et al. 1996 identified a Ty1-copia-like retroelement that was highly amplified throughout the pine genome. KVARNHEDEN et al. 1995 found evidence for six retropseudogenes in the cdc2 gene family in Picea abies. PICHERSKY 1990 notes that most postulated mechanisms of gene duplication do not result in an "instant," fully diploidized duplication. A plant's ability to self increases the likelihood of establishing the novel duplication in the diploid state, whereas in a highly outbred species this likelihood is decreased. Therefore, the process of gene amplification could produce unlinked, hemizygous nulls within the conifer genome (e.g., Figure 1).

An often overlooked advantage of mapping with forest trees is their longevity. Once the initial time and money is invested, relatively little additional effort is required to maintain this germplasm. Consequently, it is feasible to preserve these pedigrees in multiple environments and score multiple phenotypic traits over time. The genetic information pertaining to this consensus map, based on the immortalized base and qtl pedigrees, is made publicly available through the Dendrome web server and TreeGenes database for genome projects in loblolly and other conifers.

	FOOTNOTES

¹ Present address: USDA Forest Service, Southern Research Station, P.O. Box 2680, Asheville, NC 28802.
² Present address: Mendel Biotechnology, Inc., 21375 Cabot Blvd., Hayward, CA 94545.

	ACKNOWLEDGMENTS

This work has benefited from the contributions of many colleagues. In particular, the authors thank J. Cairney, P. Gustafsson, D. Harry, S. Jansson, S. Karpinski and colleagues, C. Loopstra, D. O'Malley, and R. Whetten for contributing characterized gene probes; L. Burris and N. Wheeler for maintaining pedigrees and providing germplasm; D. Bassoni, M. Devey, A. Groover, P. Hodgskiss, K. Jermstad, K. Kiehne, J. Lee, and P. Wilcox for providing technical and/or analytical support; C. Echt, S. Young, and three anonymous reviewers for providing helpful editorial comments. This research was supported by U.S. Department of Agriculture/National Research Initiative Plant Genome Grant 96-35300-3719.

Manuscript received July 22, 1998; Accepted for publication September 28, 1998.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AHUJA, M. R., M. E. DEVEY, A. T. GROOVER, K. D. JERMSTAD, and D. B. NEALE, 1994  Mapped DNA probes from loblolly pine can be used for restriction fragment length polymorphism mapping in other conifers. Theor. Appl. Genet. 88:279-282.

BEAVIS, W. D. and D. GRANT, 1991  A linkage map based on information from four F₂ populations of maize (Zea mays L.). Theor. Appl. Genet. 82:636-644.

BENNETZEN, J. L. and M. FREELING, 1993  Grasses as a single genetic system: genome composition, collinearity and compatibility. Trends Genet. 9:259-261[Medline].

BERNATZKY, R. and S. D. TANKSLEY, 1986  Majority of random cDNA clones correspond to single loci in the tomato genome. Mol. Gen. Genet. 203:8-14.

BISHOP, Y. M., S. E. FIENBURG and P. W. HOLLAND, 1975 Discrete Multivariate Analysis: Theory and Practice, pp. 492–494. MIT Press, Cambridge, MA.

BISHOP, D. T., C. CANNINGS, M. SKOLNICK and J. A. WILLIAMSON, 1983 The number of polymorphic DNA clones required to map the human genome, pp. 181–200 in Statistical Analysis of DNA Sequence Data, edited by B. S. WEIR. Marcel-Dekker, New York.

CONKLE, M. T., P. D. HODGSKISS, L. B. NUNNALLY and S. C. HUNTER, 1982 Starch gel electrophoresis of conifer seeds: a laboratory manual. Gen. Tech. Rep. PSW-64. Pacific Southwest Forest and Range Experiment Station, USDA Forest Service, Berkeley, CA.

DEVEY, M. E., K. D. JERMSTAD, C. G. TAUER, and D. B. NEALE, 1991  Inheritance of RFLP loci in a loblolly pine three-generation pedigree. Theor. Appl. Genet. 83:238-242.

DEVEY, M. E., T. A. FIDDLER, B.-H. LIU, S. J. KNAPP, and D. B. NEALE, 1994  An RFLP linkage map for loblolly pine based on a three-generation outbred pedigree. Theor. Appl. Genet. 88:273-278.

DEVEY, M. E., A. DELFINO-MIX, B. B. KINLOCH, and D. B. NEALE, 1995  Efficient mapping of a gene for resistance to white pine blister rust in sugar pine. Proc. Natl. Acad. Sci. USA 92:2066-2070[Abstract/Free Full Text].

DEVEY, M. E., J. C. BELL, D. N. SMITH, D. B. NEALE, and G. F. MORAN, 1996a  A genetic linkage map for Pinus radiata based on RFLP, RAPD, and microsatellite markers. Theor. Appl. Genet. 92:673-679.

DEVEY, M. E., M. M. SEWELL and D. B. NEALE, 1996b A comparison of loblolly and radiata pine genomes using RFLP markers, pp. 478–480 in Tree Improvement for Sustainable Tropical Forestry, edited by M. J. DIETERS, A. C. MATHESON, D. G. NIKLES, C. E. HARWOOD and S. M. WALKER. Queensland Forestry Research Institute, Gympie, Queensland, Australia.

DREWRY, A., 1988  The G-banded karyotype of Pinus resinosa Ait. Silvae Genet. 37:218-221.

DROUIN, G. and G. A. DOVER, 1987  A plant processed pseudogene. Nature 328:557-558.

ECHT, C. S. and C. D. NELSON, 1997  Linkage mapping and genome length in eastern white pine (Pinus strobus L.). Theor. Appl. Genet. 94:1031-1037.

FRANKLIN, E. C., 1970 Survey of mutant forms and inbreeding depression in species of the family Pinaceae. Research Paper SE-61. Southeastern Forest Experiment Station, Asheville, NC.

GENTZBITTEL, L., F. VEAR, Y.-X. ZHANG, A. BERVILLÉ, and P. NICOLAS, 1995  Development of a consensus linkage RFLP map of cultivated sunflower (Helianthus annus L.). Theor. Appl. Genet. 90:1079-1086.

GERBER, S. and F. RODOLPHE, 1994  An estimation of the genome length of maritime pine (Pinus pinaster Ait). Theor. Appl. Genet. 88:289-292.

GROOVER, A., M. DEVEY, J. LEE, R. MEGRAW, and T. MITCHELL-OLDS et al., 1994  Identification of quantitative trait loci influencing wood specific gravity in an outbred pedigree of loblolly pine. Genetics 138:1293-1300[Abstract].

GROOVER, A. T., C. G. WILLIAMS, M. E. DEVEY, J. M. LEE, and D. B. NEALE, 1995  Sex-related differences in meiotic recombination frequency in Pinus taeda.. J. Hered. 86:157-158[Free Full Text].

HAMRICK, J. L., and M. J. GODT, 1990 Allozyme diversity in plant species, pp. 43–63 in Plant Population Genetics, Breeding, and Genetic Resources, edited by A. H. D. BROWN, M. T. CLEGG, A. L. KAHLER and B. S. WEIR. Sinauer Associates, Sunderland, MA.

HAUGE, B. M., S. M. HANLEY, S. CARTINHOUR, J. M. CHERRY, and H. M. GOODMAN et al., 1993  An integrated genetic/RFLP map of the Arabidopsis thaliana genome. Plant J. 3:745-754.

HELENTJARIS, T., D. WEBER, and S. WRIGHT, 1988  Identification of the genomic locations of duplicate nucleotide sequences in maize by analysis of restriction fragment length polymorphisms. Genetics 118:353-363[Abstract/Free Full Text].

HULBERT, S. H., T. W. ILOTT, E. J. LEGG, S. E. LINCOLN, and R. W. MICHELMORE, 1988  Genetic analysis of the fungus, Bremia lactucae, using restriction fragment length polymorphisms. Genetics 120:947-958[Abstract/Free Full Text].

JERMSTAD, K. D., A. M. REEM, J. R. HENIFIN, N. C. WHEELER, and D. B. NEALE, 1994  Inheritance of restriction fragment length polymorphisms and random amplified polymorphic DNAs in coastal Douglas fir. Theor. Appl. Genet. 89:758-766.

KAMM, A., R. L. DOUDRICK, J. S. HESLOP-HARRISON, and T. SCHMIDT, 1996  The genomic and physical organization of Ty1-copia-like sequences as a component of large genomes in Pinus elliottii var. elliottii and other gymnosperms. Proc. Natl. Acad. Sci. USA 93:2708-2713[Abstract/Free Full Text].

KIANIAN, S. F. and C. F. QUIROS, 1992  Generation of a Brassica oleracea composite RFLP map: linkage arrangements among various populations and evolutionary implications. Theor. Appl. Genet. 84:544-554.

KINLAW, C. S. and D. B. NEALE, 1997  Complex gene families in pine genomes. Trends Plant Sci. 2:356-359.

KINLAW, C. S., S. M. GERTTULA, and M. C. CARTER, 1994  Lipid transfer protein of loblolly pine are members of a complex gene family. Plant Mol. Biol. 26:1213-1216[Medline].

KVARNHEDEN, A., K. TANDRE, and P. ENGSTRÖM, 1995  A cdc2 homologue and closely related processed retropseudogenes from Norway spruce. Plant Mol. Biol. 27:391-403[Medline].

LANDER, E. S., P. GREEN, J. ABRAHAMSON, A. BARLOW, and M. J. DALY et al., 1987  MAPMAKER: an interactive computer package for constructing primary genetic linkage maps of experimental and natural populations. Genomics 1:174-181[Medline].

LIU, Y.-G., N. MITSUKAWA, C. LISTER, C. DEAN, and R. F. WHITTIER, 1996  Isolation and mapping of a new set of 129 RFLP markers in Arabidopsis thaliana using recombinant inbred lines. Plant J. 10:733-736[Medline].

MANLY, K. F., 1993  A Macintosh program for storage and analysis of experimental genetic mapping data. Mamm. Genome 4:303-313[Medline].

MEHRA, P. N. and T. N. KHOSHOO, 1956  Cytology of conifers. Int. J. Genet. 54:165-185.

MIROV, M. T., 1967 The Genus Pinus. The Ronald Press Company, New York.

MITTON, J. B., 1994  Molecular approaches to population biology. Annu. Rev. Ecol. Syst. 25:45-69.

MORAN, G. F., J. C. BELL, and A. J. HILLIKER, 1983  Greater meiotic recombination in male vs. female gametes in Pinus radiata. J. Hered. 74:62[Abstract/Free Full Text].

NEALE, D. B., J. C. WEBER, and W. T. ADAMS, 1984  Inheritance of needle tissue isozymes in Douglas fir. Can. J. Genet. Cytol. 26:459-468.

NELSON, C. D., W. L. NANCE, and R. L. DOUDRICK, 1993  A partial genetic linkage map of slash pine (Pinus elliottii Engelm. var. elliottii) based on random amplified polymorphic DNAs. Theor. Appl. Genet. 87:145-151.

NELSON, C. D., T. L. KUBISIAK, M. STINE, and W. L. NANCE, 1994  A genetic linkage map of longleaf pine (Pinus palustris Mill.) based on random amplified polymorphic DNAs. J. Hered. 85:433-439[Abstract/Free Full Text].

O'MALLEY, D. M., S. PORTER, and R. R. SEDEROFF, 1992  Purification, characterization, and cloning of cinnamyl alcohol dehydrogenase in loblolly pine (Pinus taeda L.). Plant Physiol. 98:1364-1371[Abstract/Free Full Text].

PEDERICK, L. A., 1970  Chromosome relationships between Pinus species. Silvae Genet. 19:171-180.

PERRY, D. J. and G. R. FURNIER, 1996  Pinus banksiana has at least seven expressed alcohol dehydrogenase genes in two linked groups. Proc. Natl. Acad. Sci. USA. 93:13020-13023[Abstract/Free Full Text].

PICHERSKY, E., 1990  Nomad DNA—a model for movement and duplication of DNA sequences in plant genomes. Plant Mol. Biol. 15:437-448[Medline].

PLOMION, C., D. M. O'MALLEY, and C.-E. DUREL, 1995  Genomic analysis in maritime pine (Pinus pinaster): comparison of two RAPD maps using selfed and open-pollinated seeds of the same individual. Theor. Appl. Genet. 90:1028-1034.

QI, X., P. STAM, and P. LINDHOUT, 1996  Comparison and integration of four barley genetic maps. Genome 39:379-394.

SEWELL, M. M., and D. B. NEALE, 1998 A consensus map for loblolly pine. Dendrome database, http://dendrome.ucdavis.edu/Data/.

STAM, P., 1993  Construction of integrated genetic linkage maps by means of a new computer package: JOINMAP. Plant J. 3:739-744.

TANKSLEY, S. D., M. W. GANAL, J. P. PRINCE, M. C. DE VICENTE, and M. W. BONIERBALE et al., 1992  High density molecular linkage maps of the tomato and potato genomes. Genetics 132:1141-1160[Abstract].

TULSIERAM, L. K., J. C. GLAUBITZ, G. KISS, and J. E. CARLSON, 1992  Single tree genetic linkage mapping in conifers using haploid DNA from megagametophytes. Bio/Technology 10:686-690[Medline].

WAKAMIYA, I., R. J. NEWTON, J. S. JOHNSTON, and H. J. PRICE, 1993  Genome size and environmental factors in the genus Pinus.. Am. J. Bot. 80:1235-1241.

This article has been cited by other articles:

				T. Yin, S. P. DiFazio, L. E. Gunter, X. Zhang, M. M. Sewell, S. A. Woolbright, G. J. Allan, C. T. Kelleher, C. J. Douglas, M. Wang, et al. Genome structure and emerging evidence of an incipient sex chromosome in Populus Genome Res., March 1, 2008; 18(3): 422 - 430. [Abstract] [Full Text] [PDF]

				C. Bouza, M. Hermida, B. G. Pardo, C. Fernandez, G. G. Fortes, J. Castro, L. Sanchez, P. Presa, M. Perez, A. Sanjuan, et al. A Microsatellite Genetic Map of the Turbot (Scophthalmus maximus) Genetics, December 1, 2007; 177(4): 2457 - 2467. [Abstract] [Full Text] [PDF]

				L. Costantini, M. S. Grando, S. Feingold, S. Ulanovsky, N. Mejia, P. Hinrichsen, A. Doligez, P. This, J. A. Cabezas, and J. M. Martinez-Zapater Generation of a Common Set of Mapping Markers to Assist Table Grape Breeding Am. J. Enol. Vitic., March 1, 2007; 58(1): 102 - 111. [Abstract] [Full Text] [PDF]

				M. Sekino and M. Hara Linkage Maps for the Pacific Abalone (Genus Haliotis) Based on Microsatellite DNA Markers Genetics, February 1, 2007; 175(2): 945 - 958. [Abstract] [Full Text] [PDF]

				W. Kai, K. Kikuchi, M. Fujita, H. Suetake, A. Fujiwara, Y. Yoshiura, M. Ototake, B. Venkatesh, K. Miyaki, and Y. Suzuki A Genetic Linkage Map for the Tiger Pufferfish, Takifugu rubripes Genetics, September 1, 2005; 171(1): 227 - 238. [Abstract] [Full Text] [PDF]

				K. V. Krutovsky, M. Troggio, G. R. Brown, K. D. Jermstad, and D. B. Neale Comparative Mapping in the Pinaceae Genetics, September 1, 2004; 168(1): 447 - 461. [Abstract] [Full Text] [PDF]

				I. V. Yap, D. Schneider, J. Kleinberg, D. Matthews, S. Cartinhour, and S. R. McCouch A Graph-Theoretic Approach to Comparing and Integrating Genetic, Physical and Sequence-Based Maps Genetics, December 1, 2003; 165(4): 2235 - 2247. [Abstract] [Full Text] [PDF]

				N. Tani, T. Takahashi, H. Iwata, Y. Mukai, T. Ujino-Ihara, A. Matsumoto, K. Yoshimura, H. Yoshimaru, M. Murai, K. Nagasaka, et al. A Consensus Linkage Map for Sugi (Cryptomeria japonica) From Two Pedigrees, Based on Microsatellites and Expressed Sequence Tags Genetics, November 1, 2003; 165(3): 1551 - 1568. [Abstract] [Full Text] [PDF]

				G. R. Brown, D. L. Bassoni, G. P. Gill, J. R. Fontana, N. C. Wheeler, R. A. Megraw, M. F. Davis, M. M. Sewell, G. A. Tuskan, and D. B. Neale Identification of Quantitative Trait Loci Influencing Wood Property Traits in Loblolly Pine (Pinus taeda L.). III. QTL Verification and Candidate Gene Mapping Genetics, August 1, 2003; 164(4): 1537 - 1546. [Abstract] [Full Text] [PDF]

				T. Lenormand The Evolution of Sex Dimorphism in Recombination Genetics, February 1, 2003; 163(2): 811 - 822. [Abstract] [Full Text] [PDF]

				V. Dvornyk, A. Sirvio, M. Mikkonen, and O. Savolainen Low Nucleotide Diversity at the pal1 Locus in the Widely Distributed Pinus sylvestris Mol. Biol. Evol., February 1, 2002; 19(2): 179 - 188. [Abstract] [Full Text] [PDF]

				G. R. Brown, E. E. Kadel III, D. L. Bassoni, K. L. Kiehne, B. Temesgen, J. P. van Buijtenen, M. M. Sewell, K. A. Marshall, and D. B. Neale Anchored Reference Loci in Loblolly Pine (Pinus taeda L.) for Integrating Pine Genomics Genetics, October 1, 2001; 159(2): 799 - 809. [Abstract] [Full Text] [PDF]