The Genetic Architecture of Selection Response: Inferences From Fine-Scale Mapping of Bristle Number Quantitative Trait Loci in Drosophila melanogaster

Corresponding author: Trudy F. C. Mackay, Department of Genetics, Box 7614, North Carolina State University, Raleigh, NC 27695., trudy_mackay{at}ncsu.edu (E-mail)

Communicating editor: C. HALEY

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Quantitative trait loci (QTL) affecting responses and correlated responses to selection for abdominal and sternopleural bristle number have been mapped with high resolution to the X and third chromosomes. Advanced intercross recombinant isogenic chromosomes were constructed from high and low selection lines in an unselected inbred background, and QTL were detected using composite interval mapping and high density transposable element marker maps. We mapped a total of 26 bristle number QTL with large effects, which were in or immediately adjacent to intervals previously inferred to contain bristle number QTL on these chromosomes. The QTL contributing to response to selection for high bristle number were not the same as those contributing to response to selection for low bristle number, suggesting that distributions of allelic effects per locus may be asymmetrical. Correlated responses were more often attributable to loose linkage than pleiotropy or close linkage. Bristle number QTL mapping to the same locations have been inferred in studies with different parental strains. Of the 26 QTL, 20 mapped to locations consistent with candidate genes affecting peripheral nervous system development and/or bristle number. This facilitates determining the molecular basis of quantitative variation and allele frequencies by associating molecular variation at the candidate genes with phenotypic variation in bristle number in samples of alleles from nature.

THE bulk of phenotypic variation in populations is quantitative as opposed to qualitative; of degree rather than kind. Continuous variation for quantitative traits is caused by genetic complexity—segregating alleles at multiple loci—as well as sensitivity of the phenotypic expression of these alleles to environmental variation (FALCONER and MACKAY 1996 ). Many diseases, growth and production characters, aspects of morphology, physiology, and behavior are quantitative characters. Therefore, understanding the genetic and environmental factors affecting variation in quantitative trait phenotypes is of interest in medicine, agriculture, and evolution. Specifically, knowledge of the following properties of quantitative traits is required: the loci (quantitative trait loci, or QTL) at which mutational and segregating variation occurs; the distribution of homozygous, heterozygous, epistatic, and pleiotropic allelic effects at each locus; environmental sensitivities of QTL alleles; allele frequencies; and mutation rates.

The first stage of the genetic dissection of a quantitative trait, therefore, is to map the QTL causing genetic variation in the trait phenotype by linkage to marker loci, the segregation of which can be scored unambiguously. With the growing availability of dense polymorphic marker linkage maps, initial coarse mapping of QTL is feasible in a wide range of organisms. However, the ultimate goal of understanding the genetic architecture of quantitative traits at the level of genetic loci is most likely to be achieved using a genetically tractable organism and a model trait for which candidate genes involved in the biochemical or developmental pathways leading to the trait phenotype have been identified. One such model system is the number of sensory bristles in Drosophila melanogaster. The numbers of abdominal and sternopleural bristles have been used for over 50 years to estimate fundamental quantitative genetic parameters and to check quantitative genetic theory and were the first quantitative traits for which a comprehensive effort was made to estimate map positions of QTL (BREESE and MATHER 1957 ; THODAY 1979 ; SHRIMPTON and ROBERTSON 1988 ). Further, bristles are external sensory organs of the peripheral nervous system (PNS), and the many loci controlling PNS development that have been characterized genetically and molecularly (CAMPOS-ORTEGA 1993 ; JAN and JAN 1993 ) are candidate genes at which alleles affecting variation in bristle number in nature could segregate.

Previously, we mapped QTL causing response to divergent artificial selection for abdominal (LONG et al. 1995 ) and sternopleural (GURGANUS et al. 1999 ) bristle number from the Raleigh natural population to the X and third chromosomes. In each case, populations of recombinant isogenic (RI) chromosomes were constructed from a single generation of recombination between isogenic high and low parental line chromosomes and substituted into the low isogenic selection line backgrounds. Recombination breakpoints were determined using highly polymorphic cytological insertion sites of roo transposable elements, which provide a dense (4 cM) informative marker map. The power to detect QTL was increased by reducing the within-marker class phenotypic variation in three ways: (1) by measuring multiple individuals per RI line; (2) by mapping one chromosome at a time in a common isogenic background; and (3) by using composite interval mapping (ZENG 1994 ) to account for the segregation of chromosomally linked QTL when computing the probability that no QTL exists in each test interval. We detected two X chromosome and five chromosome 3 QTL affecting the direct response to selection for abdominal bristle number, and two X chromosome and six chromosome 3 QTL affecting the direct response to selection for sternopleural bristle number.

The results of these experiments generate questions that can only be addressed by further fine-scale mapping efforts.

1. Evidence that the estimated QTL map positions and effects might be accurate comes from overlapping map positions of QTL affecting correlated response to sternopleural (abdominal) bristle number with those determined from direct response to sternopleural (abdominal) bristle number (GURGANUS et al. 1999 ). It is nevertheless important to confirm the QTL map positions and effects by independent replication.

2. The roo transposable element insertion sites provided an even denser marker map than could be utilized, given the limited number of recombination events in a single generation. Consequently, many intervals within which the QTL mapped were quite large. To realize the ultimate goal of determining QTL locations at the level of genetic locus, much finer-scale mapping involving multiple rounds of recombination must be undertaken.

3. Many of the mapped QTL had large effects and were inferred to be at intermediate frequency in the base population. However, with large intervals, one cannot distinguish multiple linked QTL with small effects from a single QTL with a large effect, and again, further recombination is necessary to discriminate between the former, infinitesimal model of allelic effects from the Robertsonian (ROBERTSON 1967 ) exponential model.

4. Many intervals contained QTL affecting both bristle characters, and finer-scale mapping is required to distinguish close linkage from pleiotropy as the cause of these associations.

5. The initial experiments mapped the divergence between high and low QTL alleles. While this design maximizes the power to detect QTL, a consequence is that there is no information on whether the response to selection was due to both high and low alleles at each QTL or whether response at each locus was in one or the other direction. It is necessary, therefore, to map the high and low selection lines relative to an unselected background to make inferences about the distribution of allelic effects.

6. A potentially powerful technique for refining QTL map positions is by quantitative complementation tests to deficiencies and candidate loci (LONG et al. 1996 ; MACKAY and FRY 1996 ; GURGANUS et al. 1999 ). However, failure to complement in these tests can be attributable to allelism or epistasis. The latter interpretation is less likely if the alleles at other QTL on the test chromosome are not extreme, which again requires introgression of each QTL into a wild-type genetic background.

Here, we report results of fine-scale mapping of QTL affecting the response to selection for increased and decreased bristle number relative to an unselected line. This experiment utilized the same selected chromosomes as LONG et al. 1995 and GURGANUS et al. 1999 , and the map positions and effects of the QTL detected in the three studies are compared.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Drosophila stocks:
The balancer chromosomes and marker genes used are described in LINDSLEY and ZIMM 1992 . All cultures were reared in shell vials with 10 ml cornmeal-agar-molasses medium at 25°. The high and low bristle number parental lines used to establish the advanced intercross mapping populations were the same lines used by LONG et al. 1995 and GURGANUS et al. 1999 to map QTL affecting response to selection for abdominal and sternopleural bristle number, respectively, from the Raleigh base population.

• LAB: An isogenic strain derived from a line selected for low abdominal bristle number (LONG et al. 1995 ).

• H3AB: An isogenic strain in which the third chromosome derives from a line selected for high abdominal bristle number, and the first and second chromosomes are co-isogenic with LAB (LONG et al. 1995 ).

• LST1 and LST2: Two independent isogenic strains derived from a line selected for low sternopleural bristle number. LST1 was used to map chromosome 3, and LST2 to map X chromosome bristle number QTL (GURGANUS et al. 1999 ).

• H3ST: An isogenic strain in which the third chromosome derives from a line selected for high sternopleural bristle number, and the first and second chromosomes are co-isogenic with LST1 (GURGANUS et al. 1999 ).

• H1ST: An isogenic strain in which the X chromosome derives from a line selected for high sternopleural bristle number, and the second and third chromosomes are co-isogenic with LST2 (GURGANUS et al. 1999 ).

• Inbred Samarkand (Sam): This wild-type strain, derived from a standard laboratory stock by >230 generations of continuous full-sib inbreeding (LYMAN et al. 1996 ), was used as the recurrent parent into which high and low selection line alleles were introgressed.

• Sam balancer stocks: Balancer stocks of genotype Sam C(1)DX, y w f (abbreviated Sam XX below), Sam FM4, and Sam TM6B, Tb/Sb (abbreviated Sam Tb/Sb below) were constructed by substituting the balancer chromosomes into the inbred Sam background (LYMAN et al. 1996 ).

Mapping populations:
Six mapping populations—C1, HST; C1, LST; C3, HST; C3, LST; C3, HAB; and C3, LAB—were constructed by initially crossing H1ST, LST2, H3ST, LST1, H3AB, and LAB males, respectively, to inbred Sam females. Females of each population were then backcrossed to Sam males for three generations with 100–200 females mated to ~50 males per population per generation. The populations were then maintained by mass transfer with 5–10 replicate vials per population. Recombinant isogenic chromosomes were extracted during random mating generations 5–10.

To construct isogenic X chromosome lines (populations C1, HST and C1, LST), single males were crossed to Sam XX females (G1). Male progeny were crossed to (1) Sam FM4 and to (2) Sam XX females (G2), and in the following generation (G3) female FM4 heterozygote progeny of cross (1) were mated to male progeny of cross (2). The FM4 balancer chromosome was eliminated at G4 to produce an isogenic X chromosome line. To construct isogenic chromosome 3 lines (populations C3, HST; C3, LST; C3, HAB; and C3, LAB), single males were crossed to Sam Tb/Sb females (G1). Individual Tb males were backcrossed to Sam Tb/Sb females (G2), and at G3, Tb males and females were mated inter se. The Tb balancer chromosome was eliminated at G4 to produce the isogenic C3 lines.

Each of the isogenic lines was genotyped for insertion sites of roo elements (see below), and those that were recombinants between the parental selection lines and Sam were retained for further analysis (the majority of the extracted chromosomes were nonrecombinant Sam). Although these lines had been backcrossed for a total of five generations to Sam, the background genotype was further purified by backcrossing each RI chromosome 1 (3) line to Sam XX (Sam Tb/Sb) for three generations and reextracting the chromosomes as described above.

RI line genotypes:
roo insertion sites were determined by in situ hybridization of biotin-labeled roo DNA to polytene salivary chromosomes of third instar larvae according to the procedure of SHRIMPTON et al. 1986 . Phage probes containing a complete copy of roo (SCHERER et al. 1982 ) were labeled with biotinylated dATP (bio-7-dATP; Bethesda Research Laboratories, Gaithersburg, MD) by nick translation. Hybridization was detected using the Vectastain ABC kit (Vector Laboratories, Burlingame, CA) and visualized with horseradish peroxidase/3,3'-diaminobenzidine. Two high quality preparations from each RI line were scored for sites of insertion of roo elements at the level of cytological band subdivision on the standard Bridge's map (LEFEVRE 1976 ). This provides a map where the order of markers is known and recombination breakpoints in the RI lines are easily determined by the change of roo insertion sites from the parental line patterns.

RI line phenotypes:
Abdominal and sternopleural bristle number was scored on 10 males and 10 females in each of two replicate vials of all RI lines, for a total of 40 individuals scored per line. The number of bristles on the fifth abdominal sternite of males and sixth of females, and the number of bristles on the left (L) and right (R) sternopleural plates was recorded. Two sternopleural bristle traits were analyzed: the total number of bristles (L + R), and bristle number asymmetry (|L - R|/(L + R)).

Analyses of variance:
Analysis of variance (ANOVA) was used to partition the variance in bristle number and in bristle number asymmetry in each population into sources attributable to sex (S, fixed), line (L, random), S x L interaction, replicate, R(L), S x R(L), and error, E. Analyses of variance of each bristle trait were also computed for males and females separately. F-ratio tests of significance and estimates of variance components were computed using SAS GLM and VARCOMP procedures (SAS INSTITUTE 1988).

QTL mapping analysis:
The multiple opportunities for recombination afforded during the construction of the advanced intercross lines result in expanded genetic maps for each population of RI chromosomes. The map positions of the roo markers were estimated from the observed recombination frequencies between pairs of markers, r, using the Kosambi mapping function d = ¹/₄ln[(1 + 2r)/(1 - 2r)], where d is the distance between adjacent markers in morgans.

Composite interval mapping (ZENG 1993 , ZENG 1994 ), as implemented by the QTL Cartographer software (BASTEN et al. 1994 , BASTEN et al. 1997 ), was used to test the hypothesis that an interval flanked by two adjacent markers contains a QTL affecting the trait, while simultaneously controlling for the effects of chromosomally linked QTL by multiple regression on markers outside the test interval. These analyses were conducted on least-squares line means separately for each trait and sex. The conditioning markers were chosen for each analysis by forward stepwise regression. A conditioning window of 15 cM was used, such that only markers 15 cM away from the markers flanking the test interval were included in the model. The likelihood-ratio (LR) test statistic is -2 ln(L₀/L₁), where L₀/L₁ is the ratio of the likelihood under the null hypothesis (there is no QTL in the interval) to the alternative hypothesis (there is a QTL in the interval). The test statistic at a genomic location is distributed as ² with 2 d.f. under the null hypothesis, and was evaluated every centimorgan. Two significance levels were used to infer the presence of QTL. First, empirical distributions of LR test statistics under the null hypothesis of no association between any of the intervals and trait values were obtained for each analysis by randomly permuting the trait data 1000 times and calculating the maximum LR statistic across all intervals for each permutation. LR statistics from the original data that were exceeded by the permutation maximum LR statistics <50 times are significant at = 0.05 under the null hypothesis. Second, the Bonferroni correction for multiple tests, in which critical values of the LR test statistic exceeding ² [2, 0.05/n], where n is the number of independent tests per mapping population, was used as a less stringent significance threshold. The number of independent tests per chromosome was estimated as (C/50 + 1), where C is the total estimated map length of the "expanded" genetic map.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Mapping populations and bristle number phenotypes:
Six advanced intercross populations of RI lines were established by repeated crossing of X chromosomes selected for high and low sternopleural bristle number and third chromosomes divergently selected for abdominal and sternopleural bristle number to an unselected inbred strain, Sam. A total of 266 RI lines were scored for abdominal and sternopleural bristle number phenotypes. The numbers of RI lines, mean bristle numbers, and estimated magnitudes of genetic (V_L) and environmental (V_R) variance, as well as the genetic correlations of bristle number between the sexes, are given for each mapping population in Table 1. There was highly significant genetic variance between each of the selected chromosomes and Sam in both sexes and for the two bristle traits in all populations. The genetic correlation of abdominal (sternopleural) bristle number between males and females was significantly different from unity in 5 (3) of the six mapping populations, suggesting some QTL affecting bristle number have sex-specific effects. There was significant among-line variation for asymmetry of sternopleural bristle number in only one analysis: females of the C3, HAB population (data not shown).

Marker genotypes:
The cytogenetic insertion sites of roo transposable elements were determined for each of the RI lines. Markers were informative provided at least one recombination event was observed between them. The informative markers and their map positions in centimorgans, estimated from observed recombination frequencies using the Kosambi map function, are given for each population in Table 2. All genetic maps were expanded relative to the standard map (LINDSLEY and ZIMM 1992 ) as expected given multiple opportunities for recombination. There were 26 (22) informative markers for the C1, HST (LST) population, for an average interval size of 14.3 (11.4) cM on the expanded map. The corresponding interval sizes on the standard map are 2.56 cM for C1, HST and 3.08 cM for C1, LST; i.e., the X chromosome maps were expanded by factors of 5.6 and 3.7, respectively. There were 39, 30, 36, and 30 informative markers for, respectively, C3 populations HST, LST, HAB, and LAB. The average interval size for the C3 populations was 9.67 cM on the expanded map, which corresponds to 3.12 cM on the standard map, and an average map expansion by a factor of 3.1.

QTL mapping:
QTL affecting abdominal and sternopleural bristle number in each sex were mapped by linkage to molecular markers in each population using a composite interval mapping method (ZENG 1993 , ZENG 1994 ). Briefly, the likelihood that a QTL is present in a test interval is evaluated relative to the null hypothesis (no QTL in the interval), while simultaneously accounting for the segregation of chromosomally linked QTL by fitting as cofactors additional markers that explain variation in the trait, situated 15 cM away from the test interval. The conditioning markers were chosen by forward step-wise regression. The markers used as cofactors were not necessarily the same in the different analyses and varied in number from 2 to 10. The results of these analyses are summarized in Table 3 and Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6. QTL are reported in Table 3 at two significance levels: those for which the LR test statistic exceeds the permutation threshold, which is the more stringent criterion, and those for which the LR test statistics exceed the ² [2, 0.05/n] critical value (n is the number of independent tests). Less stringent thresholds are considered because we wish to pool information across sexes, traits, and populations. The permutation thresholds varied according to the trait and sex analyzed, and are shown in the figures. The thresholds based on the number of independent tests were 10.15 for the C1, HST and C3, HAB populations; 9.88 for C3, HST, C3, LST, and C3, LAB; and 9.57 for C1, LST.

View larger version (16K): In this window In a new window Download PPT slide   Figure 1. X chromosome QTL affecting response to selection for high (top) and low (bottom) sternopleural bristle number. The likelihood ratio (LR) statistic is plotted against cytogenetic map location at the level of polytene band subdivision. Each of the numbered X chromosome polytene bands (1–20) consists of 6 subdivisions (A–F), for a total of 120 subdivisions. The solid lines are for QTL affecting male bristle number and the dotted lines are for QTL affecting female bristle number. The horizontal lines represent the critical value of the LR statistic for an experimentwise type I error rate determined by permuting the trait and marker data 1000 times.

View larger version (18K): In this window In a new window Download PPT slide   Figure 2. X chromosome QTL affecting correlated response to selection in abdominal bristle number in the high (top) and low (bottom) sternopleural bristle number selection lines. The axes and notation are as in Figure 1.

View larger version (18K): In this window In a new window Download PPT slide   Figure 3. Chromosome 3 QTL affecting response to selection for high (top) and low (bottom) sternopleural bristle number. Each of the chromosome 3 polytene bands (61–100) consists of 6 subdivisions (A–F) for a total of 240 subdivisions. Otherwise, the axes and notation are as in Figure 1.

View larger version (17K): In this window In a new window Download PPT slide   Figure 4. Chromosome 3 QTL affecting correlated response to selection in abdominal bristle number in the high (top) and low (bottom) sternopleural bristle number selection lines. The axes and notation are as in Figure 3.

View larger version (21K): In this window In a new window Download PPT slide   Figure 5. Chromosome 3 QTL affecting response to selection for high (top) and low (bottom) abdominal bristle number. The axes and notation are as in Figure 3.

View larger version (18K): In this window In a new window Download PPT slide   Figure 6. Chromosome 3 QTL affecting correlated response to selection in sternopleural bristle number in the high (top) and low (bottom) abdominal bristle number selection lines. The axes and notation are as in Figure 3.

Within each population, we have estimates of map positions and effects of QTL affecting two bristle traits in both sexes. As the QTL map positions from the different analyses often coincide, we can use these analyses to deduce a minimum number of bristle number QTL for each population. We inferred that there was statistical support for the presence of a QTL within a population if (1) the LR test statistic exceeded the permutation threshold in any one of the four analyses or (2) the LR test statistic exceeded the ² threshold corrected for multiple independent tests in at least two separate analyses. Multiple QTL in a region were inferred if two or more significant LR peaks were separated by LR scores below the threshold value; otherwise a single QTL was inferred at the location of the largest significant LR statistic. [This procedure is somewhat ad hoc. Multiple trait composite interval mapping (JIANG and ZENG 1995 ) combines information from two or more traits (or the same trait in both sexes) to estimate QTL map positions and main and interaction QTL effects, and is expected to be more precise than single trait analysis. However, software applications for this method are not yet publicly available.]

For ease of reference, we designate the bristle number QTL below as "Bn," followed in parentheses by the cytological location of the marker nearest the peak LR statistic. The terminology is for convenience only and is not meant to convey the impression that the QTL have been mapped to the level of genetic locus; hence, the QTL names are not italicized as is conventional for Drosophila gene names. All effects are relative to Sam.

C1, HST: At least three QTL were detected. There are two LR peaks at the tip of the X chromosome in all analyses. The first maps to the 1A–1C interval and is associated with increased sternopleural bristle number in males and females, increased abdominal bristle number in females, and decreased abdominal bristle number in males. Bn(2A) decreases both bristle traits in males and females, and Bn(10B) is associated with increased sternopleural and abdominal bristle number in males only. There was suggestive evidence for two additional QTL with female-specific effects on abdominal bristle number in this population.

C1, LST: There is at least one QTL at the tip of the X chromosome associated with decreased sternopleural and abdominal bristle number in both sexes. In all analyses the LR test statistics displayed two peaks, one at 0.01 or 1.01 cM (2B), and one at 5.28 or 6.28 cM (between 2B and 2C). Although this region may be genetically complex, we refer to a single QTL as Bn(2B, 2C). Two QTL in this population reached the suggestive significance threshold in females.

C3, HST: There were four QTL affecting sternopleural bristle number in this population. Bn(61B) increased sternopleural bristle number in both sexes. The remaining three sternopleural bristle QTL were significant in females only; Bn(67E) and Bn(96F) were associated with increased, and Bn(91B) was associated with decreased bristle number. Only one QTL affecting abdominal bristle number was detected, Bn(93D); this QTL was associated with increased bristle number in males. There was suggestive evidence for the presence of an additional QTL affecting increased sternopleural bristle number in males [Bn(65D)].

C3, LST: Six QTL were detected in this population. Bn(67A), Bn(70E), and Bn(86B) were associated with decreased sternopleural bristle number in both sexes; the latter two QTL were also associated with a significant increase in male abdominal bristle number. Bn(91F) was significantly associated with decreased sternopleural bristle number in females. Bn(89E) and Bn(88E) were associated with decreased abdominal bristle number, the former in both sexes and the latter in females. There was suggestive evidence for one further QTL, Bn(97C), affecting abdominal bristle number in males.

C3, HAB: At least three QTL were detected with effects on abdominal bristle number. Bn(86D) and Bn(94B) increased abdominal bristle number in both sexes, and Bn(98E) increased abdominal bristle number in females. According to the decision rule that multiple QTL are inferred in a region if the LR test statistic drops below the significance threshold and then rises above it again, there appear to be three closely linked QTL affecting male sternopleural bristle number at the tip of chromosome 3. However, the magnitude of the fluctuation in LR scores is trivial. To be conservative, we postulate a single QTL, Bn(61A, 62D), in this region. There were formally two other sternopleural bristle number QTL in this population, Bn(89A), affecting male bristle number, and Bn(89F), affecting female bristle number. However, their proximity and similarity of effects in both sexes, as well as the form of the likelihood profiles, suggest that there is only one QTL in this region, Bn(89A, 89F).

C3, LAB: Two QTL decreased abdominal bristle number significantly in both sexes, Bn(72E) and Bn(92C). In addition, three abdominal bristle QTL exceeded the permutation threshold in females only, one of which decreased [Bn(64B)], and two of which increased [Bn(66C) and Bn(67E)] abdominal bristle number. There was suggestive evidence for two more female abdominal bristle QTL, at 61E and 94EF. At least three QTL affecting sternopleural bristle number were detected. Bn(62D) affected this trait in males, and Bn(68C) in both sexes. The third QTL covered a large pericentromeric region and appeared to have two peaks in the likelihood profile in both males and females (although the evidence for a female QTL in this region is suggestive only). One peak is at 77B and the other at 84B. Although this is a large physical distance, the restriction of recombination near the centromere results in a shrinkage of the genetic distance on the expanded map to only 3.3 cM. As the estimated effects are similar at both likelihood peaks, we adopt the conservative interpretation that there is only one QTL affecting sternopleural bristle number in this region, Bn(77B, 84B).

To tally the total number of QTL detected and to infer whether response to divergent selection is caused by alleles at the same or different loci, it is necessary to merge the QTL maps for the different populations. A QTL was judged to be the same in two or more populations if the map positions coincided with the same marker in those instances where the marker was present in both populations or if the map positions relative to flanking markers matched. All QTL exceeding the lower significance threshold were included in this consideration. Those that were detected at the lower threshold in only a single analysis were considered to be spurious. However, a suggestive QTL in one population that coincided with at least one suggestive QTL in another population, or with QTL that were formally significant as judged by the permutation tests, were included in the final total. The results are given in Table 4.

In total, we have mapped 26 bristle number QTL, 4 on the X and 22 on chromosome 3. Four QTL were inferred to be present in two populations, and 2 were detected in three populations. The QTL alleles fall into five categories, based on the sign of the effect and trait affected:

1. QTL alleles that affect the selected trait in the direction of selection and are thus associated with response to selection.

2. QTL alleles that affect the selected trait in the opposite direction to selection. Such QTL may occur on the selected chromosome if they are closely linked to a selected QTL; they also contribute to selection response, albeit in an undesired manner. The two sternopleural bristle QTL with large and opposite effects at the tip of the X chromosome in the HST population may well be an example of alleles in close repulsion linkage in the base population—the markers corresponding to the peak LR statistics, 1A and 1E, are in a region of very restricted recombination and are only 0.3 map units apart on the standard map (LINDSLEY and ZIMM 1992 ). QTL alleles were included in this category if they were <15 cM from the putative selected QTL on the standard map.

3. Correlated response to selection can be attributed to QTL alleles that contribute to direct response to selection but that also have pleiotropic effects on the correlated character.

4. Correlated selection response can also occur if QTL alleles affecting the correlated trait are linked to QTL affecting the directly selected trait. We inferred correlated responses were attributable to pleiotropy or very close linkage if the QTL affecting the directly selected and correlated trait mapped to identical locations, and to loose linkage if the QTL affecting the correlated trait mapped within 15 cM of the directly selected QTL.

5. The final category includes alleles with effects that do not fall into any of the previous categories. This includes QTL alleles in a selection line with effects on the selected trait in the opposite direction to selection but that are not obviously linked to selected QTL. This could occur if the allele has beneficial pleiotropic fitness effects, if it has an epistatic effect in the opposite direction to its main effect, if the effect is small, or indeed if the QTL is a false positive. QTL alleles with "wrong" signs may also have originated from Sam. Each of the 34 QTL alleles is assigned to one or more of these categories in Table 4.

Three of the four X chromosome QTL affected the selected trait, sternopleural bristle number, in the direction of selection. Two of these QTL mapped to the high chromosome and one on the low selected chromosome. Fifteen QTL alleles were associated with response to direct selection on chromosome 3: four each for the high and low sternopleural chromosomes, and three for the high and four for the low abdominal chromosomes. Perhaps surprisingly, the QTL alleles associated with direct response to selection for high bristle number were totally different from those associated with response to selection for low bristle number. Several QTL alleles affected the selected trait in the opposite direction to selection. In addition to the QTL at the tip of the X chromosome mentioned above, one QTL in the high sternopleural bristle line and three QTL in the low sternopleural bristle line fell into this category. They are all closely linked to selected QTL but could also be Sam alleles.

X chromosome QTL affecting correlated responses in abdominal bristle number mapped to the same location as those affecting direct selection response and can be inferred to be attributable to pleiotropy or close linkage. Six chromosome 3 QTL were deduced to be associated with correlated responses in abdominal bristle number, one on the high and five on the low sternopleural chromosome. Of these QTL, two mapped to the same location as QTL affecting direct selection response, and the others were within 15 cM of a directly selected QTL. Similarly, six chromosome 3 QTL were hypothesized to be associated with correlated responses in sternopleural bristle number, one on the high and five on the low selected chromosome. None mapped to the same location as QTL with effects in the direction of selection, but were within 15 map units of selected QTL. Thus, correlated responses from pleiotropic effects of directly selected QTL (or close linkage to these QTL) are less common (5) than correlated responses attributable to linkage of QTL affecting the correlated trait with QTL affecting the directly selected trait (10). The effects of two QTL alleles could not be interpreted as arising from direct or correlated selection response.

Composite interval mapping analyses were also performed for asymmetry in sternopleural bristle number. Two QTL for asymmetry were detected in females of the C3, HAB population, for which there was significant among-line variation in asymmetry. The QTL mapped to 61E (LR = 9.70) and 89F (LR = 9.53); both are nearly significant at the permutation threshold of 9.78. Although there was no among-line variation in asymmetry in any of the other populations, a QTL affecting asymmetry was mapped between 11E and 12C in the C1, HST population. The LR statistic for the QTL was 13.59, which exceeds the permutation threshold of 12.67.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Consistency:
Because QTL mapping requires very large sample sizes and there are serious problems with multiple testing in the context of genome scans, it is always necessary to confirm the map positions of QTL by independent replication. Previously, QTL affecting response to selection for abdominal bristle number and correlated response in sternopleural bristle number were mapped on the C3, HAB and C3, LAB chromosomes, relative to each other, by LONG et al. 1995 . Similarly, QTL affecting divergent response to selection for sternopleural bristle number and correlated response in abdominal bristle number were mapped on the C1, HST, C1, LST, C3, HST, and C3, LST chromosomes by GURGANUS et al. 1999 . To what extent do the map positions of QTL detected in this study coincide with those identified previously?

The abdominal bristle QTL of LONG et al. 1995 were located in intervals 61A–64C, 66A–67C, 75C–85E, 89D–92E, and 96B–96F. In this study, QTL affecting response to selection for abdominal bristle number Bn(61A, 62D), Bn(64B), Bn(66C), Bn(67A-E), Bn(72E), Bn(86D), Bn(92C), and Bn(94EF, 97C) map at, or adjacent to, the first four of these intervals. Further, QTL Bn(94B) and Bn(98E) flank the final interval. LONG et al. 1995 mapped QTL affecting correlated response in sternopleural bristle number to the 61A–64C and 89D–92E intervals. We mapped at least two QTL in and near the first region affecting sternopleural bristle number [Bn(61A, 62D) and Bn(64E)] and two in the latter [Bn(88E, 89A) and Bn(89E-F)]. Given that a slightly different composite interval mapping analysis was used by LONG et al. 1995 —all markers >10 cM away from the test interval were included in the multiple regression model, rather than those significant by forward stepwise regression, and map distances were taken from the standard map, rather than estimated from the data—the agreement between the two studies is excellent.

The composite interval mapping analysis of GURGANUS et al. 1999 detected a major factor affecting response to selection for sternopleural bristle number and correlated response in abdominal bristle number at the tip of the X chromosome (1A–3E), corresponding to our QTL Bn(1A–C), Bn(2A), and Bn(2B, 2C). A QTL affecting sternopleural and abdominal bristle number was detected at 9A–12C; this corresponds to Bn(10B–D). GURGANUS et al. 1999 mapped a QTL affecting female abdominal bristle number to the 5D–8E interval. In this analysis, a QTL affecting female abdominal bristle number at 8E was significant on the basis of the ² threshold but did not exceed the permutation threshold and was not included in the final tabulation. The correspondence between the two data sets is thus excellent.

In the analyses of GURGANUS et al. 1999 , there were six chromosome 3 QTL affecting direct response to selection for sternopleural bristle number and three chromosome 3 QTL affecting correlated response in abdominal bristle number. The most likely positions of the sternopleural bristle QTL were between 61D–62D, 65F–69F, 70E–75B, 87F–89A, 89A–89E, and 98F–100C, and of the abdominal bristle QTL, between 65F–70E, 70E–75B, and 89E–90F. Several of our factors affecting response to selection for sternopleural bristle number [Bn(61A, 62D), Bn(64E, 65D), Bn(67A), Bn(67E, 69F), and Bn(70E)] and correlated response in abdominal bristle number [Bn(70E), Bn(88E, 89A), and Bn(89E–F)] map within or adjacent to these intervals. Although additional QTL affecting direct and correlated selection response not detected by GURGANUS et al. 1999 were mapped in this study, overall there is good agreement between the two experiments.

Magnitude of effects:
Some of the QTL detected in the studies of LONG et al. 1995 and GURGANUS et al. 1999 had large effects. The mean effect of the five chromosome 3 abdominal bristle QTL was 2.4 bristles in males and 3.5 bristles in females, with ranges from 1.4–3.6 and 2.4–5.6 in males and females, respectively (LONG et al. 1995 ). Similarly, the mean effect of X and third chromosome sternopleural bristle QTL was 1.1 in males and females, with a range of 0.18–1.6 in males and 0.29–1.5 in females (GURGANUS et al. 1999 ). However, the intervals in which the QTL with large effects resided were rather wide, leaving unresolved the issue of whether a few factors with large effects or multiple linked factors with smaller effects accounted for the selection response.

We have clarified here that the minimum number of factors accounting for selection response in abdominal bristle number on chromosome 3 has grown from 5 to 10. The mean absolute value of abdominal bristle QTL effects has concomitantly dropped to 2.0 in males and 2.2 in females, with ranges of 1.1–3.2 in males and 1.4–4.7 in females. The minimum number of factors on chromosomes 1 and 3 accounting for selection response in sternopleural bristle number has risen from 2 and 6 to, respectively, 4 and 9. However, the mean absolute value of sternopleural bristle QTL effects has increased to 2.5 in males and 2.7 in females, with ranges from 1.6–4.2 and 1.4–4.8 in males and females, respectively. The effects of the QTL detected here were thus of the same magnitude, if not larger, as those estimated previously. The solution to the apparent paradox is that many of our mapped factors had effects opposite to the direction of selection; presumably these factors were on the selected chromosome as a consequence of hitchhiking along with the selected loci. Finer-scale mapping of bristle number QTL does not result in the detection of more QTL with smaller and smaller effects: the distribution of QTL effects is more close to exponential than infinitesimal.

There was good agreement between the total divergence in bristle number of parental selected chromosomes and the sum of effects of mapped QTL in previous analyses. The divergence in abdominal bristle number between the H3AB and LAB and strains was 13.7 bristles, and the sum of the abdominal bristle QTL effects mapped by LONG et al. 1995 was 14.8 bristles. Similarly, the divergence in sternopleural bristle number between the H1ST and LST2 strains was 2.3 bristles and between the H3ST and LST1 strains was 7.7 bristles; the sum of sternopleural bristle QTL effects detected by GURGANUS et al. 1999 was 1.2 bristles for X chromosome QTL and 6.5 bristles for chromosome 3 QTL. The summed effects of bristle number QTL mapped in this study agree less well with the divergences between the original parental lines. The total divergence between high and low chromosome 3 abdominal bristle QTL is only 10 bristles, nearly 4 bristles less than that between the parental lines. On the other hand, the total divergence between high and low X and chromosome 3 sternopleural bristle QTL—5.9 and 10.4 bristles—exceeds that between the parental lines by 3.6 and 2.7 bristles, respectively. A plausible explanation for these observations is epistasis: the previous estimates of parental line divergence and QTL effects were made in the low selection line background, whereas effects were estimated here in a wild-type background. If this explanation is true, the epistatic effects are in opposite directions for the two bristle traits, with high abdominal bristle alleles enhanced and high sternopleural bristle alleles diminished in the background of the low selection line.

Correlated effects:
Previously, it was found that QTL affecting correlated response to selection mapped to a subset of the same intervals in which factors affecting direct response to selection were mapped. However, the intervals were often large, and consequently it was not possible to interpret the cause of correlated responses as due to pleiotropy or linkage. Here we have shown that two-thirds of the QTL associated with correlated selection responses were linked to the QTL associated with direct selection response, and one-third were associated with the same markers as those affecting direct responses. Pleiotropy can be presumed to be the cause of the latter correlated responses pending further finer scale mapping.

Distribution of allelic effects per locus:
In no case did we observe that the QTL associated with response to selection for high sternopleural or abdominal bristle number were the same as those associated with response to selection for low bristle number. Therefore, we can infer that the distribution of allelic effects at loci affecting selection response is not such that a high- and low-effect allele segregates at each locus at intermediate frequency. If the distribution of allelic effects is symmetrical, then we must infer that high and low alleles of large effect are rare, so the probability of sampling the high allele in the high line and the low allele in the low line is very small. However, the results of experiments in which selection limits for bristle number were compared between single-pair bottleneck and larger base populations suggest that many of the alleles affecting selection response are at intermediate frequency in nature (ROBERTSON 1968 ; FRANKHAM 1980 ). It is therefore possible that the distribution of allelic effects is asymmetrical at the majority of loci contributing to selection response, such that there exists a common allele with either a high or low effect on bristle number relative to other alleles segregating at the locus.

Six of the bristle number QTL were detected in more than one population. None of the QTL alleles detected in more than one population had the same effects in each. QTL alleles in different populations affected the same trait in opposite directions or had effects on a different trait and/or sex. Either there are multiple alleles with differing effects at some of the loci affecting selection response or our description of these QTL as allelic based on their map positions and marker associations is incorrect.

If our contention is that at least some of the QTL affecting variation in bristle number in nature are alleles of loci affecting PNS development (see below), then one might expect asymmetrical distributions of effects of naturally occurring alleles and multiple alleles with varying pleiotropic effects, based on distributions of mutational allelic effects at these loci. For example, loss-of-function mutations at the achaete-scute complex, which is required to initiate sensory organ development, generally lead to loss of bristles and hairs; and loss-of-function mutations at Delta, which acts to suppress sensory organ development, generally lead to increases in bristle number (LINDSLEY and ZIMM 1992 ).

Other populations:
Are the same loci responsible for genetic variation in bristle number in different populations? This question can only be addressed by repeating the experiments described here for the Raleigh population on a number of different populations. This is a rather daunting task, but comparison of these results with earlier and more recent studies suggests that some loci may be polymorphic for alleles affecting variation in bristle number in many populations. WOLSTENHOLME and THODAY 1963 mapped two-third chromosome sternopleural bristle number QTL to the 85A–87C interval. SPICKETT and THODAY 1966 mapped four QTL affecting response to selection for high sternopleural bristle number, two at 3C and 13F on the X chromosome and two in the 67C–67D region. More recently, GURGANUS et al. 1998 mapped QTL affecting sternopleural and abdominal bristle number in a panel of recombinant inbred lines derived from two unselected inbred strains. Two X chromosome QTL associated with markers at 1B and 4C, and seven chromosome 3 QTL associated with markers at 61D, 67D, 79E, 88E, 92A, 94D, and 96F. Finally, NUZHDIN et al. 1998 mapped sternopleural bristle number QTL segregating between two wild-type inbred strains by assessing changes in neutral transposable element marker frequencies during response to divergent selection. Three QTL were detected, one on the X chromosome in the 4B–6A interval and two on chromosome 3 between 92C–94E and 97B–98E. The agreement in map positions of QTL in these studies with those reported here is striking. Many bristle number QTL mapped over a period of 35 years, using different base populations, mapping methods, and statistical analyses, coincide with positions of QTL we have mapped in the Raleigh population. Further study of loci that contribute to genetic variation in bristle number within populations and that are also polymorphic in different populations promises to be rewarding in terms of understanding the evolutionary forces maintaining quantitative genetic variation (BARTON and TURELLI 1989 ).

Candidate genes:
As noted above, one class of loci that are candidate genes at which naturally occurring variation for bristle number might segregate are the many loci that act in the progressive determination of the PNS (JAN and JAN 1993 ). Candidate genes with map positions consistent with those of the bristle number QTL are listed in Table 4. Of the 26 bristle number QTL, the map positions of 16 coincided with at least one gene known to be involved with PNS development (JAN and JAN 1993 ; SALZBERG et al. 1997 ); one coincided with a known mutation with an adult bristle number phenotype, but whose role in bristle development is unknown (LINDSLEY and ZIMM 1992 ); and three coincided with P-element insertional mutations with adult bristle number phenotypes, but which are otherwise uncharacterized (LYMAN et al. 1996 ). If QTL map to the same location as candidate genes that have been cloned and sequenced, the actual molecular basis of the mutation associated with the QTL allele and its frequency can be inferred by associating molecular polymorphisms at the candidate gene with phenotypic variation in bristle number in a large sample of alleles from nature (MACKAY and LANGLEY 1990 ; LAI et al. 1994 ; LONG et al. 1998 ).

However, some bristle number QTL with large effects, for example, Bn(10B–D), Bn(72E), and Bn(94B), do not map near known genes with bristle number phenotypes. More precise and finer-scale mapping of these loci may be feasible using overlapping deficiencies and the recombinant isogenic lines generated in the course of this study. Such studies might reveal that these loci do correspond to known loci with bristle number phenotypes; alternatively, they may not so correspond and further study may reveal novel loci discovered using quantitative genetic analysis of naturally occurring variation.

	ACKNOWLEDGMENTS

We thank Jeff Leips and two anonymous reviewers for comments on the manuscript. This work was supported by National Institutes of Health grants GM-45146 and GM-45344 to T.F.C.M., and by the W. M. Keck Foundation.

Manuscript received June 30, 1998; Accepted for publication July 21, 1999.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BARTON, N. H. and M. TURELLI, 1989  Evolutionary quantitative genetics: How little do we know? Annu. Rev. Genet. 23:337-370[Medline].

BASTEN, C. J., B. S. WEIR and Z.-B. ZENG, 1994 Zmap—a QTL cartographer, pp. 65–66 in Proceedings of the 5th World Congress on Genetics Applied to Livestock Production: Computing Strategies and Software, Vol. 22, edited by C. SMITH, J. S. GAVORA, B. BENKEL, J. CHESNAIS, W. FAIRFULL et al. The Organizing Committee, 5th World Congress on Genetics Applied to Animal Production, Guelph, Ontario, Canada.

BASTEN, C. J., B. S. WEIR and Z.-B. ZENG, 1997 QTL Cartographer: A Reference Manual and Tutorial for QTL Mapping. Department of Statistics, North Carolina State University, Raleigh, NC.

BREESE, E. L. and K. MATHER, 1957  The organization of polygenic activity within a chromosome in Drosophila. I. Hair characters. Heredity 11:373-395.

CAMPOS-ORTEGA, J. A., 1993 Early neurogenesis in Drosophila melanogaster, pp. 1091–1129 in The Development of Drosophila melanogaster, Vol. 2, edited by M. BATE and A. MARTINEZ ARIAS. Cold Spring Harbor Laboratory Press, Plainview, NY.

FALCONER, D. S., and T. F. C. MACKAY, 1996 Introduction to Quantitative Genetics, Ed. 4. Addison Wesley Longman, Harlow, Essex, United Kingdom.

FRANKHAM, R., 1980 The founder effect and response to artificial selection in Drosophila, pp. 87–90 in Selection Experiments in Laboratory and Domestic Animals, edited by A. ROBERTSON. Commonwealth Agricultural Bureaux, Slough, United Kingdom.

GURGANUS, M. C., J. D. FRY, S. V. NUZHDIN, E. G. PASYUKOVA, and R. F. LYMAN et al., 1998  Genotype-environment interaction for quantitative trait loci affecting sensory bristle number in Drosophila melanogaster.. Genetics 149:1883-1898[Abstract/Free Full Text].

GURGANUS, M. C., S. V. NUZHDIN, J. W. LEIPS, and T. F. C. MACKAY, 1999  High-resolution mapping of quantitative trait loci for sternopleural bristle number in Drosophila melanogaster.. Genetics 152:1535-1604.

JAN, Y. N., and L. Y. JAN, 1993 The peripheral nervous system, pp. 1207–1244 in The Development of Drosophila melanogaster, Vol. 2, edited by M. BATE and A. MARTINEZ ARIAS. Cold Spring Harbor Laboratory Press, Plainview, NY.

JIANG, C. and Z.-B. ZENG, 1995  Multiple trait analysis of genetic mapping for quantitative trait loci. Genetics 140:1111-1127[Abstract].

LAI, C., R. F. LYMAN, A. D. LONG, C. H. LANGLEY, and T. F. C. MACKAY, 1994  Naturally occurring variation in bristle number and DNA polymorphisms at the scabrous locus of Drosophila melanogaster.. Science 266:1697-1702[Abstract/Free Full Text].

LEFEVRE, G., 1976 A photographic representation of the polytene chromosomes of Drosophila melanogaster salivary glands, pp. 31–36 in The Genetics and Biology of Drosophila, Vol. 1a, edited by M. ASHBURNER and E. NOVITSKI. Academic Press, London.

LINDSLEY, D. L., and G. G. ZIMM, 1992 The Genome of Drosophila melanogaster. Academic Press, San Diego.

LONG, A. D., S. L. MULLANEY, L. A. REID, J. D. FRY, and C. H. LANGLEY et al., 1995  High resolution mapping of genetic factors affecting abdominal bristle number in Drosophila melanogaster.. Genetics 139:1273-1291[Abstract].

LONG, A. D., S. L. MULLANEY, T. F. C. MACKAY, and C. H. LANGLEY, 1996  Genetic interactions between naturally occurring alleles at quantitative trait loci and mutant alleles at candidate loci affecting bristle number in Drosophila melanogaster.. Genetics 144:1497-1518[Abstract].

LONG, A. D., R. F. LYMAN, C. H. LANGLEY, and T. F. C. MACKAY, 1998  Two sites in the Delta gene region contribute to naturally occurring variation in bristle number in Drosophila melanogaster.. Genetics 149:999-1017[Abstract/Free Full Text].

LYMAN, R. F., F. LAWRENCE, S. V. NUZHDIN, and T. F. C. MACKAY, 1996  Effects of single P element insertions on bristle number and viability in Drosophila melanogaster.. Genetics 143:277-292[Abstract].

MACKAY, T. F. C. and J. D. FRY, 1996  Polygenic mutation in Drosophila melanogaster: genetic interactions between selection lines and candidate quantitative trait loci. Genetics 144:671-688[Abstract].

MACKAY, T. F. C. and C. H. LANGLEY, 1990  Molecular and phenotypic variation in the achaete-scute region of Drosophila melanogaster.. Nature 348:64-66[Medline].

NUZHDIN, S. V., P. D. KEIGHTLEY, E. G. PASYUKOVA, and E. A. MOROZOVA, 1998  Mapping quantitative trait loci affecting Drosophila melanogaster sternopleural bristle number using changes of marker allele frequencies in divergently selected lines. Genet. Res. 72:79-91[Medline].

ROBERTSON, A., 1967 The nature of quantitative genetic variation, pp. 265–280 in Heritage From Mendel, edited by A. BRINK. The University of Wisconsin Press, Madison, WI.

ROBERTSON, A., 1968 The spectrum of genetic variation, pp. 5–16, in Population Biology and Evolution, edited by R. C. LEWONTIN. Syracuse University Press, Syracuse, NY.

SALZBERG, A., S. N. PROKOPENKO, Y. HE, and P. TSAI et al., 1997  P-element insertion alleles of essential genes on the third chromosome of Drosophila melanogaster: mutations affecting embryonic PNS development. Genetics 147:1723-1741[Abstract].

SAS INSTITUTE, 1988 SAS/STAT User's Guide, Release 6.03 Edition. SAS Institute, Cary, NC.

SCHERER, G., C. TSCHUDI, J. PERERA, H. DELIUS, and V. PIRROTTA, 1982  B104, a new dispersed repeated gene family in Drosophila melanogaster and its analogies with retroviruses. J. Mol. Biol. 157:435-451[Medline].

SHRIMPTON, A. E. and A. ROBERTSON, 1988  The isolation of polygenic factors controlling bristle score in Drosophila melanogaster. II. Distribution of third chromosome bristle effects within chromosome sections. Genetics 118:445-459[Abstract/Free Full Text].

SHRIMPTON, A. E., E. A. MONTGOMERY, and C. H. LANGLEY, 1986  Om mutations in Drosophila ananassae are linked to insertions of a transposable element. Genetics 114:125-135[Abstract/Free Full Text].

SPICKETT, S. G. and J. M. THODAY, 1966  Regular responses to selection. 3. Interactions between located polygenes. Genet. Res. 7:96-121[Medline].

THODAY, J. M., 1979 Polygene mapping: uses and limitations, pp. 219–233 in Quantitative Genetic Variation, edited by J. N. THOMPSON and J. M. THODAY. Academic Press, New York.

WOLSTENHOLME, D. R. and J. M. THODAY, 1963  Effects of disruptive selection. VII. A third chromosome polymorphism. Heredity 10:413-431.

ZENG, Z.-B., 1993  Theoretical basis of precision mapping of quantitative trait loci. Proc. Natl. Acad. Sci. USA 90:10972-10976[Abstract/Free Full Text].

ZENG, Z.-B., 1994  Precision mapping of quantitative trait loci. Genetics 136:1457-1468[Abstract].

This article has been cited by other articles:

				S. J. Macdonald and A. D. Long Joint Estimates of Quantitative Trait Locus Effect and Frequency Using Synthetic Recombinant Populations of Drosophila melanogaster Genetics, June 1, 2007; 176(2): 1261 - 1281. [Abstract] [Full Text] [PDF]

				H.-B. Park, L. Jacobsson, P. Wahlberg, P. B. Siegel, and L. Andersson QTL analysis of body composition and metabolic traits in an intercross between chicken lines divergently selected for growth Physiol Genomics, April 13, 2006; 25(2): 216 - 223. [Abstract] [Full Text] [PDF]

				S. J. Macdonald, T. Pastinen, and A. D. Long The Effect of Polymorphisms in the Enhancer of split Gene Complex on Bristle Number Variation in a Large Wild-Caught Cohort of Drosophila melanogaster Genetics, December 1, 2005; 171(4): 1741 - 1756. [Abstract] [Full Text] [PDF]

				R. C. Fuller, C. F. Baer, and J. Travis How and When Selection Experiments Might Actually be Useful Integr. Comp. Biol., June 1, 2005; 45(3): 391 - 404. [Abstract] [Full Text] [PDF]

E. A. Ivakine, C. J. Fox, A. D. Paterson, S. M. Mortin-Toth, A. Canty, D. S. Walton, K. Aleksa, S. Ito, and J. S. Danska Sex-Specific Effect of Insulin-Dependent Diabetes 4 on Regulation of Diabetes Pathogenesis in the Nonobese Diabetic Mouse J. Immunol., June 1, 2005; 174(11): 7129 - 7140.