Epistatic Interactions Between smell-impaired Loci in Drosophila melanogaster

Grayna M. Fedorowicz

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

Communicating editor: L. PARTRIDGE

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS AND DISCUSSION LITERATURE CITED

Odor-guided behavior is a polygenic trait determined by the concerted expression of multiple loci. Previously, P-element mutagenesis was used to identify single P[lArB] insertions, in a common isogenic background, with homozygous effects on olfactory behavior. Here, we have crossed 12 lines with these smell impaired (smi) mutations in a half-diallel design (excluding homozygous parental genotypes and reciprocal crosses) to produce all possible 66 doubly heterozygous hybrids with P[lArB] insertions at two distinct locations. The olfactory behavior of the transheterozygous progeny was measured using an assay that quantified the avoidance response to the repellent odorant benzaldehyde. There was significant variation in general combining abilities of avoidance scores among the smi mutants, indicating variation in heterozygous effects. Further, there was significant variation among specific combining abilities of each cross, indicating dependencies of heterozygous effects on the smi locus genotypes, i.e., epistasis. Significant epistatic interactions were identified for nine transheterozygote genotypes, involving 10 of the 12 smi loci. Eight of these loci form an interacting ensemble of genes that modulate expression of the behavioral phenotype. These observations illustrate the power of quantitative genetic analyses to detect subtle phenotypic effects and point to an extensive network of epistatic interactions among genes in the olfactory subgenome.

THE fundamental goal of quantitative genetics is to understand how complex traits are shaped through the interactions of multiple genes in different genetic backgrounds and under varying environmental conditions. Perhaps the most complex category of polygenic traits is represented by various forms of animal behavior. Drosophila melanogaster presents an ideal model system to study the genetic basis of behavioral quantitative traits, because mutations in highly inbred strains can be easily generated, allowing control over the segregation of many individual loci that contribute to the trait and enabling the effect of each locus to be studied independently. We have used odor-guided behavior in D. melanogaster as a model system to study the quantitative genetics of behavior.

Odor-guided behavior is of special interest, because the ability of an organism to respond to chemical signals from its environment is essential for its survival and, often, its procreation. Thus, olfactory behavior contributes to individual fitness (MACKAY et al. 1996 ). In recent years, considerable progress has been made in elucidating the molecular mechanisms that underlie odor recognition, olfactory transduction, and neural coding of olfactory information both in vertebrates (reviewed by ANHOLT 1993 ; AXEL 1995 ; BUCK 1996 ) and in invertebrate model systems, such as Caenorhabditis elegans (TROEMEL et al. 1996 ; SENGUPTA et al. 1996 ) and lobster (FADOOL and ACHE 1992 ). However, the genetic basis of variation in olfactory responsiveness and the genetic mechanisms that shape behavioral responses to odorants are still poorly understood.

Chemical mutagenesis has been used to induce mutations affecting olfactory behavior in D. melanogaster, mostly on the X chromosome (RODRIGUES and SIDDIQI 1978 ; ACEVES-PINA and QUINN 1979 ; HELFAND and CARLSON 1989 ; LILLY and CARLSON 1989 ; MCKENNA et al. 1989 ; AYER and CARLSON 1992 ; WOODARD et al. 1992 ; LILLY et al. 1994A , LILLY et al. 1994B ). This resulted in the characterization of a number of genes that encode proteins likely to participate in olfactory signal transduction in Drosophila, such as smellblind (an allele of paralytic), which encodes a voltage-gated sodium channel (RODRIGUES and SIDDIQI 1978 ; ACEVES-PINA and QUINN 1979 ; LILLY and CARLSON 1989 ; LILLY et al. 1994A , LILLY et al. 1994B ), norpA, which encodes a phospholipase C (WOODARD et al. 1992 ; RIESGO-ESCOVAR et al. 1995 ), and rdgB, which encodes a phosphatidyl inositol transfer protein (VIHTELIC et al. 1993 ). Although mutations in any of these genes cause extensive impairment of olfactory behavior, it is not clear how these genes contribute quantitatively to variation in olfactory responsiveness and how they function in the context of the genetic background, i.e., the entire olfactory subgenome. Recently, we have identified 14 loci that contribute to olfactory behavior by P -element insertional mutagenesis in an isogenic strain (ANHOLT et al. 1996 ). Identification of these loci, designated smell impaired (smi), was achieved using statistical and quantitative genetic analysis of measurements of olfactory behavior. These analyses are capable of detecting small phenotypic effects with a resolution limited only by sample size.

As the smi loci have similar phenotypes, they are likely to be functionally related and participate in common physiological and/or developmental pathways that shape olfactory responsiveness. One genetic method for identifying and ordering genes in functionally interacting groups is to screen for mutations at unlinked loci that enhance or suppress the mutant effects of a known member of the pathway (GARCIA-BELLIDO 1981 ). Epistatic interactions between such genes can be deduced by examining the phenotypes of the one- and two-locus genotypes. For independent loci, the phenotypes of the two-locus genotypes are the sum of the single-locus phenotypes; i.e., the loci act additively. Departures from strict additivity indicate epistatic, or interacting, loci. A simple test for interaction that can be used for recessive mutations with large, qualitative effects that have similar loss-of-function phenotypes and that therefore affect a common process, is to examine the phenotypes of double mutant heterozygotes. Epistasis is evident when the double heterozygote has the same loss-of-function phenotype as the single homozygous mutations as a result of combined haploinsufficiency of function. This approach and variants of it have been used to identify epistatic interactions and to identify new loci that modify mutant phenotypes of other loci (BOTAS et al. 1982 ; BELOTE et al. 1985 ; KENNISON and RUSSELL 1987 ; HOMYK and EMERSON 1988 ; TRICOIRE 1988 ; DAMBLY-CHAUDIERE et al. 1988 ).

Detecting interactions between mutations with quantitative effects is more difficult, because the mutations are not usually completely recessive (MACKAY et al. 1992 ; LYMAN et al. 1996 ). Further, the background genotype needs to be controlled to enable small phenotypic effects to be perceived and to ensure any interactions are due to epistasis between the mutations of interest, and are not confounding nonadditive interactions among alleles segregating between the background genotypes in which the mutations were induced. The smi mutations are in a common isogenic background and therefore can be used to detect epistasis. We have generated all possible double heterozygous hybrids among 12 independent smi mutations that appear amenable to molecular characterization in a diallel cross design (GRIFFING 1956 ), which is the quantitative genetic analogue of the transheterozygote test for epistasis. This approach is based on the assumption that reduced expression of two independent P[lArB]-tagged smi genes in double heterozygotic offspring may result in quantitative failure to complement (MACKAY and FRY 1996 ; LONG et al. 1996 ) if these genes interact.

Significant epistatic interactions were identified for nine transheterozygote genotypes, involving 10 of the 12 smi loci. Interactions between eight of these loci show evidence of a web of mutually interactive genes, the coordinated expression of which modulates the behavioral phenotype. These findings illustrate the power of quantitative genetic analyses to detect subtle phenotypic effects and indicate that phenotypic determination of odor-guided behavior in D. melanogaster depends quantitatively on an extensive network of genetic interactions.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS AND DISCUSSION LITERATURE CITED

Generation of transheterozygous P[lArB] insert lines:
The parental lines used to generate double mutant heterozygotes were 12 homozygous smi lines obtained by P-element mutagenesis of the isogenic Samarkand; ry⁵⁰⁶ strain: smi21F, smi26D, smi27E, smi28E, smi35A, smi45E, smi51A, smi60E, smi61A, smi65A, smi97B, and smi98B (ANHOLT et al. 1996 ). The mutations were named according to the cytological insertion sites of the P elements. The 12 smi lines were crossed in a half-diallel design (excluding homozygous parental lines and reciprocal crosses) to produce all 66 possible combinations of F₁ transheterozygous offspring with two P elements at different loci. Crosses were initiated at a density of five females of smi line i and five males of smi line j (i j) in plastic culture vials. All animals were reared at 25° on agar-yeast-molasses medium.

Behavioral assay:
To quantify odor-guided behavior we used the simple, rapid, and highly reproducible "dipstick" assay, described previously (ANHOLT et al. 1996 ). This assay was chosen because it has several advantages over other commonly used assays. The Y-maze assay, developed by RODRIGUES and SIDDIQI 1978 , is better suited for measurements of attraction and odor discrimination than repulsion and is laborious for large behavioral screens. The olfactory jump assay, described by MCKENNA et al. 1989 , was in our hands unreliable, because in contrast to Canton-S flies for which this assay was developed, flies of both our inbred Samarkand strain and of substitution lines containing chromosomes from natural populations (MACKAY et al. 1996 ) seldom jumped in response to repellent odorants. The "dipstick assay" used in this study and in previous studies (ANHOLT et al. 1996 ; MACKAY et al. 1996 ) is a simple, rapid, and highly reproducible statistical sampling assay that quantifies odor-guided behavior with a resolution limited only by sample size and, hence, can detect subtle olfactory impairments (ANHOLT et al. 1996 ). Previously, this assay has led to the identification of 14 novel smi loci. In addition, known mutants, such as smellblind (RODRIGUES and SIDDIQI 1978 ; ACEVES-PINA and QUINN 1979 ; LILLY and CARLSON 1989 ; LILLY et al. 1994A , LILLY et al. 1994B ), are immediately apparent and readily quantifiable in this assay (ANHOLT et al. 1996 ).

After 2–4 hr of starvation, 2–10-day post-eclosion transheterozygous progeny were tested for responsiveness to benzaldehyde, a repellent odorant, exactly as described by ANHOLT et al. 1996 . Briefly, one replicate assay consisted of a single-sex group of five individuals in a test vial. The animals were exposed to 1% benzaldehyde (v/v) introduced on a cotton wool swab, and the number of flies migrating to a compartment remote from the odor source was measured at 5-sec intervals, from 15 to 60 sec after introduction of the odor source. The "avoidance score" of the replicate is the average of these 10 counts, giving a possible range of avoidance scores between 0 (all flies in the compartment near the odor source for the entire assay period) and 5 (all flies in the compartment away from the odor source for the entire assay period). For each of the 66 crosses, 10 replicate avoidance score estimates were obtained for each sex, for a total of 20 replicates (100 individual flies) per double heterozygote genotype, and a total sample size of 6600 animals.

Statistical analyses:
The avoidance scores of transheterozygous genotypes were analyzed by two-way analysis of variance (ANOVA), with Genotype and Sex the fixed cross-classified main effects. Sums of squares were partitioned into sources (degrees of freedom) attributable to Genotype (65), Sex (1), Genotype x Sex interaction (65), and Error (1188). As this is a fixed effects model, the error mean square was used as the denominator for all F-ratio tests of significance. To analyze epistatic effects between smi loci, we could not simply compare the responses of double heterozygotes with the single heterozygotes of smi lines with Sam, because the effect of P[lArB] insert copy number (2 vs. 1) could be confounding. Rather, the correct control for this analysis is measurement of the deviation from the average of all other transheterozygotes with the two single inserts being compared. Thus, the general combining ability (GC A) of a mutation is its average avoidance score as a transheterozygote with all other mutations, expressed as the deviation from the overall mean (SPRAGUE and TATUM 1942 ), and is an estimate of the average heterozygous effect of the mutation relative to the heterozygous effects of the other mutations. The specific combining ability (SCA) of a transheterozygous genotype is the difference between the observed avoidance score of the genotype, x_ij (where i and j denote two different smi mutations), and the score expected from the sum of the corresponding GC As of mutants i and j. The sums of squares due to Genotype and Genotype x Sex were further partitioned into sources of variation (degrees of freedom) attributable to GC A (11), SC A (54), GC A x Sex (11), and SC A x Sex (54). SC A effects are due to variation in heterozygous effects that depend on the genetic background with respect to other smi mutations and can only be caused by epistatic interactions.

This fixed effects half-diallel corresponds to Method 4, Model I of GRIFFING 1956 . Consequently, the GC A for each smi mutant was estimated as

(1)

FALCONER and MACKAY 1996

GRIFFING 1956

(2)

The significance of the overall GC A, SC A, GC A x Sex, and SC A x Sex effects was tested using an F variance ratio test statistic with the error mean square as the denominator. Standard errors of individual GC A and SC A effects were computed according to the formulae given by GRIFFING 1956 . Analyses of variance and tests of significance were calculated using SA S procedures (SA S INSTITUTE, INC. 1988 ), and GC A and SC A sums of squares were computed using the diallel cross analysis program of SCHAFFER and USANIS 1969 .

	RESULTS AND DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS AND DISCUSSION LITERATURE CITED

The effects of 12 P -element insertional mutations with homozygous effects on olfactory behavior were evaluated in all possible double heterozygote combinations, in a half-diallel design. The mean avoidance responses to benzaldehyde, averaged over sexes, are shown for each of the 66 transheterozygote genotypes in Table 1. The analysis of variance of these data is given in Table 2. The differences in mean avoidance responses among the heterozygous genotypes were highly significant (P = 0.0001). There was also significant sexual dimorphism in avoidance response to benzaldehyde, averaged over all genotypes (P = 0.0009), with a mean male avoidance score of 4.1 and a mean female score of 3.9. Sexual dimorphism for olfactory avoidance response has been observed previously for homozygous P -element insertional mutations (ANHOLT et al. 1996 ) and among a sample of isogenic X and third chromosomes extracted from a natural population and substituted into the same inbred strain used for P -element mutagenesis (MACKAY et al. 1996 ). Interestingly, both the homozygous P -element insertions and the naturally occurring alleles affecting olfactory behavior had very large genotype x sex interaction effects, indicating that there was variation in the magnitude of the sex dimorphism of effects among the homozygous genotypes. However, the genotype x sex interaction was not significant for the double heterozygote genotypes; therefore, the sex-specific effects observed previously are on average recessive.

Variation among the transheterozygous genotypes can arise from two sources: variation in mean heterozygous effects of the different mutations, and variation from epistatic interactions. Because all P -element insertions are in the same inbred strain, all genetic variation among the genotypes is attributable to one of these two sources, with no confounding effects contributed by the background genotype. Classical diallel cross analysis enables us to separate heterozygous from epistatic effects by partitioning the variation among double heterozygous genotypes into their general (GC A) and specific (SC A) combining abilities. As mentioned above, the GC A of a mutation is an estimate of its mean heterozygous effect in the background of each of the other mutations. Estimates of the GC A of each smi mutation, expressed as deviations from the overall mean of the population of heterozygous genotypes, are given in Table 1. For comparison, also given in Table 1 are the mean avoidance scores of each smi mutation, at the same concentration of odorant used to assess transheterozygote olfactory behavior (HOM; ANHOLT et al. 1996 ). All homozygous smi mutations have reduced avoidance scores relative to the transheterozygotes. Therefore, negative GC A effects reflect lower mean heterozygous avoidance scores and a more mutant heterozygous phenotype; conversely, positive GC A effects reflect higher than average mean heterozygous scores and a more wild-type phenotype. This variation in GC A among the smi mutations is highly significant (P < 0.0001, Table 3), and from this we can infer that all the smi mutations are not completely recessive.

The overall mean avoidance score of the transheterozygous genotypes, 3.99 ± 0.08 (Table 1), is significantly higher than that of the Sam; ry ⁵⁰⁶ strain, 3.65 ± 0.08 (ANHOLT et al. 1996 ). Technically, this could be interpreted as overdominance for olfactory behavior. However, the selectable visible marker used in this system of P -element mutagenesis is ry⁺, and there is a concern that this marker has a direct effect on fitness and other quantitative traits relative to the ry^- mutant background of the control strain (LYMAN et al. 1996 ). For this reason, we cannot use these data to estimate d, the value of the heterozygote expressed as a deviation from the mean mutant and control strain value (FALCONER and MACKAY 1996 ) for each smi mutation. However, we can estimate the average degree of dominance of the smi mutations from the slope of the regression b of GC A on homozygous avoidance score of smi mutations, where each homozygous score is expressed as a deviation from the overall homozygous mutant mean. The estimate of the average degree of dominance k is 2(b - 0.5) (MACKAY et al. 1992 ; LYMAN et al. 1996 ), where k ranges from -1 (completely recessive mutations) through 0 (strict additivity) to 1 (completely dominant mutations). For these data, b = 0.197 (0.01 < P < 0.05) and k = -0.605. On average, the smi mutations are partially recessive.

The SC A of a pair of mutations reflects the extent to which the mean avoidance score of the double heterozygote, expressed as a deviation from the mean of the total population of heterozygous genotypes, departs from that expected given the sum of the GC As of the two mutant parents. Typically, diallel crosses are made among inbred lines that each vary at a number of loci affecting the measured trait, and significant SC A effects can only be attributed to nonadditive interactions in general, including dominance and epistasis (FALCONER and MACKAY 1996 ). However, in this experimental design the genetic background has been standardized, and SC A interactions can only result from epistasis, i.e., variation in heterozygous effects of a smi mutation depending on the genetic background with respect to other smi mutations.

We observed highly significant SC A effects (P = 0.0025, Table 3) for olfactory avoidance among the transheterozygote genotypes. This observation is not a scale effect. The effect of SC A was also highly significant if log, square root, and square transformations are applied to the data (data not shown). This suggests that epistatic interactions among loci affecting olfactory behavior are very common, because we have sampled only a small fraction of the total number of possible genotypes at 12 loci, each with two alleles ( = 3.8%). To determine which interacting mutations contributed to the overall variation in SC A, we determined for which transheterozygote lines SC A effects are significantly different from zero. The results are given in Table 4. Nine transheterozygous crosses reveal statistically significant epistatic interactions between smi loci. In addition, the smi98B/smi60E transheterozygote has an SC A value (-0.251) that is nearly formally significant (P = 0.063). In five of the nine statistically significant cases, the difference between the observed and expected avoidance scores (SC A) is negative; i.e., the avoidance response of the double heterozygote is more mutant than would be expected given the average degrees of dominance of both parents. In four cases, the SC A estimates were positive, indicating better olfactory responses of the hybrid offspring than expected from the average heterozygous effects of parental mutations. It should be noted that all of the transheterozygotes show avoidance scores within wild-type range, i.e., complementation, but it is the quantitative analysis of the degree of complementation that reveals epistatic effects. The negative and positive interactions are quantitative genetic analogues of mutations that enhance or suppress, respectively, the effects of other mutations affecting the same phenotype.

The observed epistatic effects are quite large; the mean of the absolute values of significant SC A effects is 0.34. This value is of the same magnitude as the mean of the absolute values of significant GC A effects (0.23), and is one-half of the environmental standard deviation. However, it is clear that these "large" quantitative effects are very subtle in absolute terms and cannot be discerned without quantitative genetic analysis of the phenotypes, or in variable genetic backgrounds. The magnitude of the epistatic effects are not necessarily correlated with the size of the homozygous mutant effects. smi loci with relatively small effects on olfactory behavior of homozygotes, e.g., smi21F and smi45E (ANHOLT et al. 1996 ), produce large effects in double heterozygous progeny. We cannot, however, determine to what extent each locus of an interacting pair of loci contributes to the observed epistatic effect. Furthermore, we do not know to what extent the P[lArB] insertion limits the expression of the gene it affects. We predict, therefore, that epistatic effects will be stronger in double heterozygotes that contain null mutations, such as deletions, at the smi loci.

The pattern of interactions observed is interesting. Of the 12 smi loci, 10 interact with at least one other. Epistatic interactions between eight smi loci can be represented in a simple interaction diagram (Figure 1). smi60E and smi61A interact, but are independent of the others. It is possible that smi98B interacts with smi60E (the P value of the SC A is on the borderline of formal statistical significance), which would place 11 of the 12 smi genes in two interacting groups. It is somewhat surprising that the mutation that interacts most extensively with other smi mutants, smi21F, has itself very weak homozygous effects. The mutant phenotype of this gene is only apparent at a low concentration of benzaldehyde and is strongly sexually dimorphic (only females display aberrant olfactory responses; males are not significantly different from wild type) (ANHOLT et al. 1996 ). Yet it elicits strong interactions in transheterozygotes with four of the smi mutations, and in both sexes.

View larger version (10K): In this window In a new window Download PPT slide   Figure 1. —Interaction diagram of smi loci. The and symbols indicate epistatic effects that suppress and enhance the homozygous mutant phenotype, respectively. Two loci, smi60E and smi61A, form an independent pair with a positive epistatic effect (not shown).

These loci represent only a small sample of the genes that affect olfactory behavior. The frequency with which smi lines were detected in our previous P -element mutagenesis screen indicated that ~4% of the Drosophila genome participates in shaping odor-guided behavior (ANHOLT et al. 1996 ), which corresponds to a conservative estimate of about 400 genes. Most likely, the ensemble of genes illustrated in Figure 1 is integrated into a more extensive network of interactions within the olfactory subgenome. Thus, loci that appear noninteractive, i.e., smi35A and (possibly) smi98B, and loci that interact independently of the larger ensemble, i.e., smi60E and smi61A, may prove to be part of a wider network of interacting genes once more olfactory genes are identified. Such extensive epistatic interactions between smi loci indicate that they form a complex network of genes that together shape odor-guided behavior.

In recent years, other investigators have identified olfactory mutants in D. melanogaster, mostly with mutations located on the X chromosome (VIHTELIC et al. 1993 ; WOODARD et al. 1992 ; RIESGO-ESCOVAR et al. 1995 ; LILLY and CARLSON 1989 ; LILLY et al. 1994A , LILLY et al. 1994B ), but epistatic interactions among them and their effects on phenotypic variation have not been assessed. Our observations suggest that olfactory genes identified on the X chromosome might also interact within functional genetic networks and these possible interactions could also include the smi loci described here. However, different genetic backgrounds may render the detection of such epistatic effects more difficult than detection of epistasis among smi genes in a coisogenic background.

Because each of the smi genes used in this study is tagged by a P element, it will, in principle, be possible in future studies to characterize their expression products and to obtain an understanding of the molecular basis for the observed genetic interactions. Moreover, our ability to use coisogenic P[lArB]-insertion lines for the characterization of networks of interacting genes in the olfactory subgenome will enable the future identification of new olfactory genes by virtue of epistatic interactions with known smi genes. Thus, these experiments pave the road for the use of quantitative genetic analysis of subtle phenotypes as a tool for targeted gene discovery.

	FOOTNOTES

¹ Present address: Department of Biology, Utah State University, Logan, UT 84332-5305.

	ACKNOWLEDGMENTS

We thank DR. RICHARD F. LYMAN for assistance in analyzing the data and J. BRANT HACKETT for technical assistance. This work was supported by grants from the National Institutes of Health (DC-02485, GM-45344, and GM-45146) and the U.S. Army Research Office (DAAH04-96-1-0096).

Manuscript received July 14, 1997; Accepted for publication January 5, 1998.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS AND DISCUSSION LITERATURE CITED

ACEVES-PIÑA, E. and W. QUINN, 1979  Learning in normal and mutant Drosophila larvae. Science 206:93-96[Abstract/Free Full Text].

ANHOLT, R. R. H., 1993  Molecular neurobiology of olfaction. Crit. Rev. Neurobiol. 7:1-22[Medline].

ANHOLT, R. R. H., R. F. LYMAN, and T. F. C. MACKAY, 1996  Effects of single P -element insertions on olfactory behavior in Drosophila melanogaster. Genetics 143:293-301[Abstract].

AXEL, R., 1995  The molecular logic of smell. Sci. Amer. 273:154-159.

AYER, R. K. and J. CARLSON, 1992  Olfactory physiology in the Drosophila antenna and maxillary palp: acj6 distinguishes two classes of odorant pathways. J. Neurobiol. 23:965-982[Medline].

BELOTE, J. M., M. B. MCKEOWN, D. J. ANDREW, T. N. SCOTT, and B. S. BAKER, 1985  Control of sexual differentiation in Drosophila melanogaster. Cold Spring Harbor Symp. Quant. Biol. 50:605-614[Abstract/Free Full Text].

BOTAS, J., J. M. DEL PRADO, and A. GARCIA-BELLÍDO, 1982  Gene-dose titration analysis in the search of trans-regulatory genes in Drosophila. EMBO J. 1:307-310[Medline].

BUCK, L. B., 1996  Information coding in the olfactory system. Annu. Rev. Neurosci. 19:517-544[Medline].

DAMBLY-CHAUDIÈRE, C., A. GHYSEN, L. Y. JAN, and Y. N. JAN, 1988  The determination of sense organs in Drosophila: interaction of scute with daughterless. Roux's Arch. Dev. Biol. 197:419-423.

FADOOL, D. A. and B. W. ACHE, 1992  Plasma membrane inositol-1,4,5-triphosphate-activated channels mediate signal transduction in lobster olfactory receptor neurons. Neuron 9:907-918[Medline].

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

GARCÍA-BELLIDO, A., 1981 The bithorax syntagma, pp. 135–148 in Advances in Genetics, Development and Evolution of Drosophila, edited by S. LAKOVAARA. Plenum Press, NY.

GRIFFING, B., 1956  Concept of general and specific combining ability in relation to diallel crossing systems. Aust. J. Biol. Sci. 9:463-493.

HELFAND, S. L. and J. CARLSON, 1989  Isolation and characterization of an olfactory mutant in Drosophila with a chemically specific defect. Proc. Natl. Acad. Sci. USA 86:2908-2912[Abstract/Free Full Text].

HOMYK, T. and C. P. EMERSON, 1988  Functional interactions between unlinked muscle genes within haploinsufficient regions of the Drosophila genome. Genetics 119:105-121[Abstract/Free Full Text].

KENNISON, J. A. and M. A. RUSSELL, 1987  Dosage-dependent modifiers of homeotic mutations in Drosophila melanogaster. Genetics 116:75-86[Abstract/Free Full Text].

LILLY, M. and J. CARLSON, 1989  Smellblind : a gene required for Drosophila olfaction. Genetics 124:293-302[Abstract].

LILLY, M., R. KREBER, B. GANETZKY, and J. R. CARLSON, 1994a  Evidence that the Drosophila olfactory mutant smellblind defines a novel class of sodium channel mutation. Genetics 136:1087-1096[Abstract].

LILLY, M., J. RIESGO-ESCOVAR, and J. CARLSON, 1994b  Developmental analysis of the smellblind mutants: evidence for the role of sodium channels in Drosophila development. Dev. Biol. 162:1-8[Medline].

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].

LYMAN, R. F., F. LAWRENCE, S. NUZDHIN, 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., R. F. LYMAN, and M. S. JACKSON, 1992  Effects of P -element insertions on quantitative traits in Drosophila melanogaster.. Genetics 130:315-332[Abstract].

MACKAY, T. F. C., J. B. HACKETT, R. F. LYMAN, M. L. WAYNE, and R. R. H. ANHOLT, 1996  Quantitative genetic variation of odor-guided behavior in a natural population of Drosophila melanogaster. Genetics 144:727-735[Abstract].

MCKENNA, M., P. MONTE, S. HELFAND, C. WOODARD, and J. CARLSON, 1989  A simple chemosensory response in Drosophila and the isolation of acj mutants in which it is affected. Proc. Natl. Acad. Sci. USA 86:8118-8122[Abstract/Free Full Text].

RIESGO-ESCOVAR, J., D. RAHA, and J. R. CARLSON, 1995  Requirement for a phospholipase C in odor response: overlap between olfaction and vision in Drosophila. Proc. Natl. Acad. Sci. USA 92:2864-2868[Abstract/Free Full Text].

RODRIGUES, V. and O. SIDDIQI, 1978  Genetic analysis of chemosensory pathway. Proc. Ind. Acad. Sci. 87B:147-160.

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

SCHAFFER, H. E., and R. A. USANIS, 1969 General least squares analysis of diallel experiments: a computer program—DIALL. Genetics Department, North Carolina State University.

SENGUPTA, P., J. C. CHOU, and C. I. BARGMANN, 1996  odr-10 encodes a seven transmembrane domain olfactory receptor required for responses to the odorant diacetyl. Cell 84:899-909[Medline].

SPRAGUE, G. F. and L. A. TATUM, 1942  General vs. specific combining ability in single crosses of corn. J. Amer. Soc. Agron. 34:923-932.

TRICOIRE, H., 1988  Dominant maternal interactions with Drosophila segmentation genes. Roux's Arch. Dev. Biol. 197:115-123.

TROEMEL, E. R., J. H. CHOU, N. D. DWYER, H. A. COLBERT, and C. I. BARGMANN, 1996  Divergent seven transmembrane receptors are candidate chemosensory receptors in C. elegans. Cell 83:207-218.

VIHTELIC, TH., M. GOEBL, S. MILLIGAN, J. E. O'TOUSA, and D. R. HYDE, 1993  Localization of Drosophila retinal degeneration B, a membrane-associated phosphatidylinositol transfer protein. J. Cell Biol. 122:1013-1022[Abstract/Free Full Text].

WOODARD, C., E. ALCORTA, and J. CARLSON, 1992  The rdgB gene in Drosophila: a link between vision and olfaction. J. Neurogenetics 8:17-32[Medline].

This article has been cited by other articles:

				D. Sambandan, M. A. Carbone, R. R. H. Anholt, and T. F. C. Mackay Phenotypic Plasticity and Genotype by Environment Interaction for Olfactory Behavior in Drosophila melanogaster Genetics, June 1, 2008; 179(2): 1079 - 1088. [Abstract] [Full Text] [PDF]

				S. M. Rollmann, A. Yamamoto, T. Goossens, L. Zwarts, Z. Callaerts-Vegh, P. Callaerts, K. Norga, T. F. C. Mackay, and R. R. H. Anholt The Early Developmental Gene Semaphorin 5c Contributes to Olfactory Behavior in Adult Drosophila Genetics, June 1, 2007; 176(2): 947 - 956. [Abstract] [Full Text] [PDF]

				D. Sambandan, A. Yamamoto, J.-J. Fanara, T. F. C. Mackay, and R. R. H. Anholt Dynamic Genetic Interactions Determine Odor-Guided Behavior in Drosophila melanogaster Genetics, November 1, 2006; 174(3): 1349 - 1363. [Abstract] [Full Text] [PDF]

				K. S. Kendler and R. J. Greenspan The Nature of Genetic Influences on Behavior: Lessons From "Simpler" Organisms Am J Psychiatry, October 1, 2006; 163(10): 1683 - 1694. [Abstract] [Full Text] [PDF]

				C. C. Spencer and D. E. L. Promislow Age-Specific Changes in Epistatic Effects on Mortality Rate in Drosophila melanogaster J. Hered., September 1, 2005; 96(5): 513 - 521. [Abstract] [Full Text] [PDF]

				B. van Swinderen and R. J. Greenspan Flexibility in a Gene Network Affecting a Simple Behavior in Drosophila melanogaster Genetics, April 1, 2005; 169(4): 2151 - 2163. [Abstract] [Full Text] [PDF]

				N. H. Kulkarni, A. H. Yamamoto, K. O. Robinson, T. F. C. Mackay, and R. R. H. Anholt The DSC1 Channel, Encoded by the smi60E Locus, Contributes to Odor-Guided Behavior in Drosophila melanogaster Genetics, August 1, 2002; 161(4): 1507 - 1516. [Abstract] [Full Text] [PDF]

				R. R.H. Anholt, J. J. Fanara, G. M. Fedorowicz, I. Ganguly, N. H. Kulkarni, T. F.C. Mackay, and S. M. Rollmann Functional Genomics of Odor-guided Behavior in Drosophila melanogaster Chem Senses, February 1, 2001; 26(2): 215 - 221. [Abstract] [Full Text] [PDF]