Female Site-Specific Transposase-Induced Recombination: A High-Efficiency Method for Fine Mapping Mutations on the X Chromosome in Drosophila

Corresponding author: Jeffrey M. Marcus, 107 Dorsheimer, SUNY, Buffalo, NY 14260., jeffmarc{at}buffalo.edu (E-mail)

Communicating editor: K. V. ANDERSON

	ABSTRACT

TOP ABSTRACT MATHEMATICAL JUSTIFICATION FOR... MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

P-element transposons in the Drosophila germline mobilize only in the presence of the appropriate transposase enzyme. Sometimes, instead of mobilizing completely, P elements will undergo site-specific recombination with the homologous chromosome. Site-specific recombination is the basis for male recombination mapping, since the male germline does not normally undergo recombination. Site-specific recombination also takes place in females, but this has been difficult to study because of the obscuring effects of meiotic recombination. Using map functions, I demonstrate that it is possible to employ female site-specific transposase-induced recombination (FaSSTIR) to map loci on the X chromosome and predict that FaSSTIR mapping should be more efficient than meiotic mapping over short genetic intervals. Both FaSSTIR mapping and meiotic mapping were used to fine map the crossveinless locus on the X chromosome. Both techniques identified the same 10-kb interval as the probable location of the crossveinless mutation. Over short intervals (< 7.6 cM), FaSSTIR produces more informative recombination events than does meiotic recombination. Over longer intervals, FaSSTIR is not always more efficient than meiotic mapping, but it produces the correct gene order. FaSSTIR matches the expectations suggested by the map functions and promises to be a useful technique, particularly for mapping X-linked loci.

MEIOTIC recombination mapping is one of the cornerstones of classical genetics. In genetic model organisms, it is commonly used to order genetic loci on chromosomes, and the precise mapping of loci can greatly assist in the identification of the molecular lesion responsible for a particular phenotype of interest. Meiotic mapping provides two distinct types of data that are used in combination to create a genetic map. Recombination distances are used to estimate the lengths of intervals between genetic loci and the behavior of flanking markers is used to infer the order of the loci.

Like all techniques, recombination mapping has its limitations. In particular, it can be used only to map loci that are within 50 cM of one another, because loci that are >50 cM apart behave as if they were on separate chromosomes. To map genetic loci that are >50 cM apart, other loci with intermediate map positions must first be identified and the markers of interest mapped with respect to these loci. A further limitation is that for genetic loci that are very close together, large numbers of progeny must be scored to detect rare recombination events between those markers, making the fine mapping of genetic loci a labor-intensive undertaking.

Recently, CHEN et al. 1998 devised a method to circumvent these limitations of meiotic mapping in Drosophila melanogaster, allowing for the rapid ordering of a locus of interest with respect to a series of P-element inserts at known locations along the chromosome. CHEN et al. 1998 took advantage of two unusual aspects of Drosophila biology: the general lack of meiotic recombination on all chromosomes in males (MORGAN 1912 , MORGAN 1914 ) and the ability to induce site-specific recombination at the site of P-element insertions by the addition of an external source of transposase (PRESTON and ENGELS 1996 ; PRESTON et al. 1996 ). By using site-specific recombination in male Drosophila, it is possible to construct crosses in such a way that all of the observed recombination events are associated with P-element insertion sites. The frequency of these recombination events, like the frequency of P-element transposition events, is due to characteristics of the P element itself and the nature of the DNA sequence surrounding the insertion site (BERG and SPRADLING 1991 ). In addition, because there is no recombination other than that at the site of the P-element insertion, the selection of flanking markers is simplified. The only requirements are that they be easy to score and that a marker be on either side of the locus of interest. Male recombination mapping has been used to fine map several autosomal loci (MCKIM and HAYASHI-HAGIHARA 1998 ; CHEN et al. 2000 ; PAI et al. 2000 ; CHU et al. 2001 ; CHUNG et al. 2001 ; LEE and TREISMAN 2001 ).

Male recombination mapping also has limitations. It generally cannot be used to map loci on the sex chromosomes because males are the hemizygous sex in Drosophila (MORGAN 1910 ) and their X and Y chromosomes are without true homologs with which to recombine (CHEN et al. 1998 ). Researchers wishing to map loci on the X chromosome (which makes up 25% of the Drosophila genome) have therefore had to use traditional meiotic mapping in females, which becomes increasingly labor intensive as the distance between the locus of interest and the markers being used to map the locus decreases. Presented here is a new technique called female site-specific transposase-induced recombination (FaSSTIR) mapping, which combines some of the benefits of male recombination mapping with the ability to map loci on the X chromosome. This technique relies on the fact that the phenomenon of site-specific recombination associated with P elements is not limited to male Drosophila and occurs in females as well (BROADHEAD et al. 1977 ; KIDWELL 1977 ). In addition, because the recombination rate between any two loci in female Drosophila is directly related to the physical distance between them on the chromosome, as increasingly shorter intervals are considered, the recombination distance approaches zero and is therefore analogous to the absence of recombination observed in males. Over sufficiently short intervals, in crosses where both meiotic and site-specific recombination are allowed to take place, site-specific recombination will be more frequent than meiotic recombination and may thus be useful in mapping closely spaced loci on the X chromosome.

As long as appropriate flanking markers are selected so that they are not lost by meiotic recombination, this technique could be used to map loci on the X chromosome and under some circumstances may be more efficient than standard meiotic mapping. Presented here are a mathematical justification for the efficacy of FaSSTIR mapping, a demonstration of the use of the technique for mapping the crossveinless (cv) locus (BRIDGES 1920 ), and a comparison of the precision and efficiency of FaSSTIR with traditional meiotic mapping.

	MATHEMATICAL JUSTIFICATION FOR FEMALE SITE-SPECIFIC TRANSPOSASE-INDUCED RECOMBINATION MAPPING

TOP ABSTRACT MATHEMATICAL JUSTIFICATION FOR... MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In Drosophila females, using a map function first suggested by HALDANE 1919 , the measured meiotic recombination frequency C_AB between two loci A and B can be expressed in terms of the map distance x_AB (in centimorgans) as

(1)

As x_AB increases from 0, C_AB increases from 0 to 0.5. Moreover, as x_AB becomes increasingly small, C_AB and x_AB converge, so that over short intervals (in which multiple crossover events are unlikely), the measured recombination frequency equals the map distance. Similarly, as x_AB approaches 0, the observed recombination frequency also approaches 0, so that

(2)

In Drosophila males the situation is somewhat different because there is no meiotic recombination in the male germline (MORGAN 1912 , 1914), so regardless of the distance between A and B (x_AB) there is no intrachromosomal recombination in males and

(3)

Site-specific recombination can be induced in males at P-element transposon insertion sites by the presence of transposase (PRESTON and ENGELS 1996 ). Male site-specific recombination occurs with a constant frequency, k_PM, that is independent of genetic distance between a locus A and a particular transposon insertion P, but instead appears to be due to the nature of the transposon itself and the characteristics of the DNA sequences flanking the transposon (BERG and SPRADLING 1991 ). Unlike meiotic recombination, site-specific recombination is essentially a binary process with no theoretical upper limit [although as an empirical matter, site-specific recombination rates >0.5 are unknown (CHEN et al. 1998 )]. Male site-specific recombination can be written as

(4)

Recombination frequencies in males are not true map distances and cannot be used to determine distances between loci. Instead, left-right positional information is obtained from the behavior of flanking markers and it is these data that are used to create a map where the order of the loci is specified, but not the distances between them (CHEN et al. 1998 ). However, since many P elements have known locations on both the physical and the DNA sequence maps, this technique can be used to map genetic loci to small precisely defined intervals between identified P-element insertions, which can facilitate the molecular characterization of genetic loci.

Site-specific recombination at the site of transposon insertions also occurs in females (KIDWELL 1977 ), but it is much more difficult to study due to obscuring effects of meiotic recombination. Both meiotic and transposase-induced recombination events can influence observed recombination frequencies in females. The effects of both of these processes can be expressed as a function of the meiotic recombination rate x_AP and the site-specific recombination rate k_PF as

(5)

It is necessary to distinguish between male site-specific recombination rates k_PM and female site-specific recombination rates k_PF because there are currently no data on the relationship between these two rates for individual transposon insertions. The generalized behavior of Equation 1, Equation 3, and Equation 5 is shown graphically in Fig 1.

View larger version (19K): In this window In a new window Download PPT slide   Figure 1. (Left axis) Theoretical predictions of the behavior of meiotic mapping (Equation 1), male recombination mapping (Equation 3), and FaSSTIR mapping (Equation 5). (Right axis) Empirical determination of the relative efficiency of meiotic mapping and FaSSTIR mapping as meiotic map distance from the locus of interest increases. Meiotic recombination rates approach 0 as distance approaches 0. Both male recombination rates and FaSSTIR rates approach nonzero constants (kPM and kPF, respectively) as distance approaches 0. As distances become large, meiotic recombination rates and FaSSTIR rates approach 50, while male recombination rates remain constant. For the comparisons of relative efficiency, meiotic recombination distance on the x-axis has been centered here on the map position of the cv locus (1-13.7). Efficiencies of 1 indicate that FaSSTIR and meiotic recombination rates are equal. Efficiencies >1 indicate FaSSTIR is more efficient, and those <1 indicate that meiotic recombination is more efficient. The greatest increase in efficiency is over the shortest intervals and generally declines as the distance from the locus of interest increases. The shaded box shows the region over which FaSSTIR was consistently more efficient than meiotic recombination, a region that encompasses 7.65 cM to the left of cv and 7.55 cM to the right of the cv locus. However, the relative increase in efficiency achieved by FaSSTIR beyond the 5-cM meiotic recombination distance to either side of the locus of interest is marginal.

In Drosophila females, as x_AB becomes increasingly small, the effects of meiotic recombination diminish, and distance-independent site-specific recombination predominates. Therefore, over sufficiently short intervals, the measured recombination rate will be due entirely to the rate of site-specific recombination and thus will be a constant k_PF so that

(6)

This observation suggests that FaSSTIR can be used to map loci over short intervals in females just as it is used to map loci over longer intervals in males (cf. Equation 4 and Equation 6). Just as in male site-specific recombination, only left-right information determined from flanking markers can be used to produce maps because FaSSTIR recombination rates are not true map distances. The choice of flanking markers is somewhat more difficult in females than in males in that they must be close enough to the loci being examined so that they are not lost by recombination. This flanking marker recombination is analogous to that encountered in typical meiotic mapping in females, which is not an issue in Drosophila male recombination mapping.

Equation 6 suggests that over short intervals and for female site-specific recombination rates greater than zero (k_PF > 0), FaSSTIR will take place more often than meiotic recombination alone. Due to this increase in potentially informative recombination events, FaSSTIR may be efficient in providing high-resolution fine-mapping data for regions of the Drosophila genome that are not amenable to male site-specific recombination mapping, particularly the X chromosome. These predictions are not predicated by the choice of map function. A similar set of equations can be derived from the alternative generalized map function of KOSAMBI 1944 and give substantially similar results to those shown here for the map function proposed by HALDANE 1919 (derivation not shown).

	MATERIALS AND METHODS

TOP ABSTRACT MATHEMATICAL JUSTIFICATION FOR... MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Drosophila stocks:
Flies were raised on standard cornmeal-molasses Drosophila medium and grown at 25° unless otherwise noted. The y¹ w¹¹¹⁸ cv¹ sn³ mapping chromosome was created by recombination. The transposase stock Sp¹/CyO; Dr¹ Delta2-3/TM6 Ubx and all P elements were obtained from the Bloomington Fly Stock Center except for EP(X)1368, EP(X)1405, EP(X)0371, and EP(X)1604, which were obtained from Exelixis. The information about cytological locations and flanking sequences of P elements is from the BERKELEY DROSOPHILA GENOME PROJECT (2002). Only homozygous viable P-element inserts with sequenced flanking regions were used. All genetic symbols not described in the text are in the Drosophila reference works (LINDSLEY and ZIMM 1992 ). P-element flanking sequences were compared to the Drosophila genomic scaffold sequence (ADAMS et al. 2000 ) by BLAST (ALTSCHUL et al. 1997 ) to determine the exact locations of the insertion sites.

Drosophila crosses:
For FaSSTIR-mapping experiments, female flies carrying the mapping chromosome y¹ w¹¹¹⁸ cv¹ sn³; +; + were mated to male flies from the transposase stock that were +; Sp¹/CyO; Dr¹ Delta2-3/TM6 Ubx. Male progeny from this stock that were y¹ w¹¹¹⁸ cv¹ sn³; Sp¹/+; Dr¹ Delta2-3/+ or y¹ w¹¹¹⁸ cv¹ sn³; CyO/+; Dr¹ Delta2-3/+ were then crossed to flies carrying the P-element chromosome [y⁺ w¹¹¹⁸ EP(X) cv⁺ sn⁺; +; +]. Virgin female progeny from this cross that were y¹ w¹¹¹⁸ cv¹ sn³/y⁺ w¹¹¹⁸ EP(X) cv⁺ sn⁺; +/+; Dr¹ Delta2-3/+ were crossed back to males carrying the mapping chromosome y¹ w¹¹¹⁸ cv¹ sn³; +; + and the progeny from this cross were scored.

For meiotic-mapping experiments, female flies carrying the mapping chromosome y¹ w¹¹¹⁸ cv¹ sn³; +; + were mated to male flies carrying the P-element chromosome y⁺ w¹¹¹⁸ EP(X) cv⁺ sn⁺; +; +. Virgin female progeny that were y¹ w¹¹¹⁸ cv¹ sn³/y⁺ w¹¹¹⁸ EP(X) cv⁺ sn⁺; +/+; +/+ were crossed to male y¹ w¹¹¹⁸ cv¹ sn³; +; + and the progeny were scored.

About five males and five females were placed in each vial. For each chosen P-element line, from 10 to 70 crosses were set up for both FaSSTIR and meiotic mapping, and progeny were scored for recombinants. Meiotic recombination events between cv and P elements close to cv are very rare, requiring greater numbers of progeny to calculate recombination frequencies and the larger number of crosses. Informative single recombinants were identified by the phenotypes produced by the flanking markers yellow and singed and by eye color produced by the mini-white construct carried by the P elements. Recombination distances and standard errors were calculated by standard methods (WEIR 1996 ). Comparisons of FaSSTIR and meiotic mapping were made by calculating the relative efficiency of the two methods by dividing the FaSSTIR recombination rate by the meiotic recombination rate for each P-element insertion. When this ratio was greater than one, FaSSTIR was more efficient than meiotic recombination. The limits of the interval in which FaSSTIR was more efficient were determined by linear interpolation.

	RESULTS

TOP ABSTRACT MATHEMATICAL JUSTIFICATION FOR... MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Scheme for mapping:
Fig 2 illustrates the schemes used for mapping a mutation using FaSSTIR and meiotic recombination. Hemizygous lethal and homozygous lethal mutations can also be mapped by FaSSTIR. Such a scheme requires the creation of double-balanced stocks, and only some recombinant progeny may be easily scored in the F₁ backcross, but is otherwise essentially similar to what is described here.

View larger version (12K): In this window In a new window Download PPT slide   Figure 2. Mapping schemes. (A) Meiotic recombination-mapping scheme. (B) FaSSTIR-mapping scheme. For the experiments described in this article, m1 corresponds to the y locus, m2 corresponds to the sn locus, g corresponds to the cv locus, w corresponds to the white (w) locus, P[w+] corresponds to P elements carrying a wild-type white genetic construct, 2-3 is the source of transposase, and D corresponds to Drop (Dr).

In the case of a hemizygous viable mutation for both FaSSTIR and meiotic recombination, two visible markers (m₁ and m₂) that flank a mutation of interest (g) are selected. The chromosome containing the P element is put in trans to a chromosome carrying mutation g that also carries mutant alleles of the visible markers. The transposase source (Delta2-3) for FaSSTIR is provided by another chromosome carrying a dominant visible marker (D). The cross must be conducted in a background (usually eye color w^- or ry^-) in which a visible marker associated with the P element (w⁺ or ry⁺) can also be scored to distinguish between single- and multiple-recombination events. For both types of mapping, a single-recombination event between the P element and g locus results in the g mutation cosegregating with a wild-type eye color marker and with either m₁ or m₂. The other single-recombination product can also be identified, because it carries only the other flanking marker on a g⁺ and w^- or ry^- chromosome. By identifying probable single recombinants, it is then possible to test whether g cosegregates with either m₁ or m₂.

Mapping the crossveinless gene:
The cv gene has been mapped to X chromosome region 5A8–5C2 in complementation tests with Df(1)C149 (fails to complement, breakpoints at 5A8–9 and 5C5–6) and Df(1)N73 (complements, breakpoints 5C2 and 5D5–6; SLIZYNSKA 1964 ) and so P elements that mapped to this region were initially chosen for analysis. The P-element insertion sites were mapped by in situ hybridization by the BERKELEY DROSOPHILA GENOME PROJECT (2002). Additional P elements that mapped to X chromosome divisions 1–10 were also tested to determine the genetic distances over which FaSSTIR is effective. A total of 23 P-element insertion lines, which span approximately half of the X chromosome, were tested. The body color locus yellow (y) and the bristle morphology locus singed (sn) were used as visible flanking markers. The cv¹ allele has a visible recessive phenotype and is both homozygous and hemizygous viable, so determining whether it cosegregated with y or sn was unambiguous.

For all of the P elements tested, it was possible to unambiguously map crossveinless with respect to the P element by both the meiotic-mapping and the FaSSTIR-mapping techniques. Eight P elements were found to be distal to the crossveinless locus and recombinants between the P and cv tended to produce y⁺ w P cv sn and y w cv⁺ sn⁺ recombination products. Fifteen P elements were found to be proximal to the crossveinless locus and recombinants between the P and cv tended to produce y w cv P sn⁺ and y⁺ w cv⁺ sn recombination products. Additional meiotic recombination events that removed or added flanking markers also took place on these recombinant chromosomes at low frequency. These events did not obscure the mapping signal, but do necessitate examining multiple recombinant progeny.

In almost all cases, the frequency of FaSSTIR exceeded the frequency of meiotic recombination between the P element and the crossveinless locus (Fig 3). Two exceptions (EP1395 and EP0912) had a lower frequency of FaSSTIR than of meiotic recombination, but still indicated the correct order of loci along the chromosome. It appears that FaSSTIR gives the greatest increases in mapping efficiency relative to meiotic mapping when the P element is relatively close to the locus of interest with a maximum relative efficiency of >200-fold. This increase in efficiency declines for insertion sites that are farther away, and for P elements that are separated by > 5 cM from locus of interest, FaSSTIR produces minimal increases in efficiency (Fig 1). This suggests that, as a mapping tool, FaSSTIR is most appropriate for mapping intervals within 5 cM of a locus of interest on the meiotic recombination map. For intervals >5 cM, it is probably advantageous to use meiotic mapping because it requires fewer generations to set up the appropriate mapping cross.

View larger version (17K): In this window In a new window Download PPT slide   Figure 3. Recombination frequencies for meiotic recombination and FaSSTIR mapping. Sequence position refers to the locations of the P-element insertion sites on the Drosophila genome scaffold sequence. The sequence position is numbered such that the distal end is 0 and increases proximally. , meiotic recombination frequencies; , FaSSTIR recombination frequencies. Standard errors were generally small and error bars for most points would have been inside the symbols, so they are not shown.

The data produced by different methods of mapping crossveinless are entirely congruent with each other. Deficiency mapping defined the interval 5A8–C2. Meiotic recombination mapping and FaSSTIR mapping both define a smaller interval, 5A13–B1, between P-element inserts EP0496 and EP1349 within the deficiency interval. Meiotic mapping further defines this interval as measuring 0.07 cM. By examining the insertion sites of the P elements on the Drosophila genome scaffold sequence, this interval can also be defined as being 10 kb in length. This interval contains three complete open reading frames and probably contains regulatory sequences for at least two other loci (Fig 4).

View larger version (46K): In this window In a new window Download PPT slide   Figure 4. Fine-scale map around the crossveinless locus. Deficiency mapping, meiotic mapping, and FaSSTIR mapping all produce congruent results for the location of the crossveinless locus. The hatched box shows the region defined by deficiency mapping and the solid box shows the region defined by meiotic and FaSSTIR mapping. The order of the P elements was determined by reported cytology and by comparing flanking sequences to the genomic scaffold sequence. These two methods produced entirely congruent orders of the P elements. The asterisk indicates cytological map positions inferred from the mapping data presented here; the star indicates map positions obtained from FLYBASE 2002 ; and the black cross indicates P elements with map positions, as determined by meiotic recombination frequencies, which would change the order of the insertions from that suggested by cytology and by comparisons of flanking sequence with the genomic scaffold (BERKELEY DROSOPHILA GENOME PROJECT 2002). For these insertions, linear interpolation with respect to the two nearest neighboring insertions was used to assign an adjusted map position for these elements. Also shown is the 10-kb genomic region to which the crossveinless lesion was mapped. It contains three complete open reading frames and may also contain regulatory sequences from neighboring genes.

	DISCUSSION

TOP ABSTRACT MATHEMATICAL JUSTIFICATION FOR... MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

I have shown that FaSSTIR can be used to map a locus to the same interval as meiotic recombination mapping and that, over short intervals, FaSSTIR is more efficient than meiotic recombination. Over longer intervals, the efficiency of FaSSTIR and meiotic recombination rates becomes similar. This conforms to the mathematical predictions that over short intervals FaSSTIR rates will be higher and over longer intervals FaSSTIR and meiotic recombination rates are expected to converge at 50%. Transposase-induced recombination in females was generated in 23 P-element insertion lines, and in each case the order of the insertion relative to the mutant of interest and to two flanking markers was obtained unambiguously. FaSSTIR allows some of the increases in the efficiency of mapping that have been derived from male recombination mapping on the autosome to be applied to genes on the X chromosome. The increasingly large collection of P-element insertions on the X chromosome will make this a highly useful technique for characterizing such loci.

Although FaSSTIR can be used on any chromosome with a homolog (in Drosophila, all except the Y chromosome), male recombination is likely to be an easier mapping technique for most autosomal loci because of the absence of meiotic recombination in males, which allows for greater flexibility in the selection of flanking markers. In addition, FaSSTIR may also be affected by phenomena similar to the misleading recombination events that are produced when some P-element inserts are used for male recombination mapping (CHU et al. 2001 ). The mechanism by which these misleading results are produced is unknown, but may also affect transposase-induced recombination in females. To minimize the effects of these misleading events, researchers using FaSSTIR should try to score as many recombinant flies as is practical. Finally, as shown above, FaSSTIR efficiency is similar to meiotic mapping for intervals > 7.6 cM and is only marginally more efficient over intervals >5 cM.

FaSSTIR may find application in two other areas. First, FaSSTIR, like male recombination, can be used to generate recombination events in regions of the genome where meiotic recombination is infrequent, such as near centromeres. This has the potential to greatly facilitate the recombination-dependent interval mapping of single-nucleotide polymorphism (SNP) markers relative to genes of interest (HOSKINS et al. 2001 ).

Second, FaSSTIR can be used to compare the transposase-induced site-specific recombination rates in male and female Drosophila. The germline in the two sexes is organized very differently and may present significantly different environments in which P-element transposition and recombination take place. However, if a set of P elements is mapped via male recombination, FaSSTIR, and female recombination, it will be possible to measure k_PM (Equation 4) and k_PF (Equation 6) for the same individual inserts. Comparing these rates in the context of male and female germ cell formation may provide insights into the biology of these transposon insertions that have contributed so significantly to the field of genetics.

Recently, it has been suggested that CG12410, the predicted open reading frame immediately proximal to the interval defined by these mapping experiments, encodes a homolog of the twisted gastrulation (tsg) locus and has been called twisted gastulation-2 (tsg2; ROSS et al. 2001 ). Twisted gastrulation protein regulates signaling by bone morphogenetic protein (BMP) ligands, such as decapentaplegic and glass-bottom boat [which play roles in crossvein formation (CONLEY et al. 2000 ; MARCUS 2001 )], by facilitating the binding of the BMP antagonist chordin and also by facilitating the degradation of cleaved chordin molecules (LARRAIN et al. 2001 ). In this manner, tsg can both antagonize and promote BMP signaling and it has been hypothesized that tsg2 may function in a similar manner (LARRAIN et al. 2001 ). It has also been reported that mutations in CG12410 fail to complement cv¹ (L. MARCH, personal communication to FlyBase: FBgn0000394), suggesting that this predicted open reading frame corresponds to the crossveinless locus. If these findings and the genetic mapping data from this study are both correct, this suggests further that the lesion that produced the cv¹ mutation is probably a regulatory mutation in the 5' sequence upstream of CG12410 and may be located in between the EP1349 insertion site and the 5' transcription start site of CG3160 (Fig 4). Confirmation of this interpretation will require additional experiments.

	ACKNOWLEDGMENTS

I thank my Ph.D. advisor, Fred Nijhout, and Julia Bowsher, Lou D'Amico, Laura Grunert, Armin Moczek, Yuichiro Suzuki, Tomalei Vess, and Andy Yang for useful discussions and comments on the manuscript. I also thank Rick Fehon for supporting and encouraging me in my Drosophila work. Amy Bejsovec, Celest Berg, Dan Kiehart, Hubie Amrein, and Robin Wharton provided essential early feedback on the practicality and general utility of this approach. John Mercer provided extremely helpful mathematical expertise. Heather Solari Gavilan, Jenn Genova, Sarah Hughes, Rima Kulikauskas, Dennis Lajeunesse, Sushmita Maitra, Andy McClelland, and Olga Nikiforova always made themselves available for consultation on esoteric Drosophila issues. Thanks go to Joanne Seiff and Lucy Seiff for continuing moral support. I am especially indebted to Sarah Barden and Kashmir Hill who each scored many many thousands of flies. Bloomington Stock Center and Exelixis kindly provided Drosophila stocks. This work was supported by funding from the Departments of Zoology and Biology at Duke University. J.M.M was supported by a Howard Hughes Medical Institute Predoctoral Fellowship.

Manuscript received September 13, 2002; Accepted for publication October 29, 2002.

	LITERATURE CITED

TOP ABSTRACT MATHEMATICAL JUSTIFICATION FOR... MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ADAMS, M. D., S. E. CELNIKER, R. A. HOLT, C. A. EVANS, and J. D. GOCAYNE et al., 2000  The genome sequence of Drosophila melanogaster.. Science 287:2185-2195.[Abstract/Free Full Text]

ALTSCHUL, S. F., T. L. MADDEN, A. A. SCHAFFER, J. ZHANG, and Z. ZHANG et al., 1997  Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25:3389-3402.[Abstract/Free Full Text]

BERG, C. A. and A. C. SPRADLING, 1991  Studies on the rate and site-specificity of P-element transposition. Genetics 127:515-524.[Abstract]

BERKELEY DROSOPHILA GENOME PROJECT, 2002 BDGP: the Drosophila genome database (http://www.fruitfly.org).

BRIDGES, C. B., 1920  The mutant crossveinless in Drosophila melanogaster.. Proc. Natl. Acad. Sci. USA 6:660-663.[Free Full Text]

BROADHEAD, R. S., J. F. KIDWELL, and M. G. KIDWELL, 1977  Variation of the recombination fraction in Drosophila melanogaster females. J. Hered. 68:323-326.[Free Full Text]

CHEN, B., T. CHU, E. HARMS, J. P. GERGEN, and S. STRICKLAND, 1998  Mapping of Drosophila mutations using site-specific male recombination. Genetics 149:157-163.[Abstract/Free Full Text]

CHEN, B., E. HARMS, T. Y. CHU, G. HENRION, and S. STRICKLAND, 2000  Completion of meiosis in Drosophila oocytes requires transcriptional control by Grauzone, a new zinc finger protein. Development 127:1243-1251.[Abstract]

CHU, T., G. HENRION, V. HAEGELI, and S. STRICKLAND, 2001  Cortex, a Drosophila gene required to complete oocyte meiosis, is a member of the Cdc20/fizzy protein family. Genesis 29:141-152.[Medline]

CHUNG, Y. D., J. C. ZHU, Y. G. HAN, and M. J. KERNAN, 2001  nompA encodes a PNS-specific, ZP domain protein required to connect mechanosensory dendrites to sensory structures. Neuron 29:415-428.[Medline]

CONLEY, C. A., R. SILBURN, M. A. SINGER, A. RALSTON, and D. ROHWER-NUTTER et al., 2000  Crossveinless 2 contains cysteine-rich domains and is required for high levels of BMP-like activity during the formation of the cross veins in Drosophila.. Development 127:3947-3959.[Abstract]

FLYBASE, 2002 The Drosophila genetic database (http://flybase.bio.indiana.edu).

HALDANE, J. B. S., 1919  The combination of linkage values, and the calculation of distances between the loci of linked factors. J. Genet. 8:299-309.

HOSKINS, R. A., A. C. PHAN, M. NAEEMUDDIN, F. A. MAPA, and D. A. RUDDY et al., 2001  Single nucleotide polymorphism markers for genetic mapping in Drosophila melanogaster.. Genome Res. 11:1100-1113.[Abstract/Free Full Text]

KIDWELL, M. G., 1977  Reciprocal differences in female recombination associated with hybrid dysgenesis in Drosophila melanogaster.. Genet. Res. 30:77-88.[Medline]

KOSAMBI, D. D., 1944  The estimation of map distances from recombination values. Ann. Eugen. 12:172-175.

LARRAIN, J., M. OELGESCHLAGER, N. I. KETPURA, B. REVERSADE, and L. ZAKIN et al., 2001  Proteolytic cleavage of Chordin as a switch for the dual activities of Twisted gastrulation in BMP signaling. Development 128:4439-4447.

LEE, J. D. and J. E. TREISMAN, 2001  Sightless has homology to transmembrane acyltransferases and is required to generate active Hedgehog protein. Curr. Biol. 11:1147-1152.[Medline]

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

MARCUS, J. M., 2001  The development and evolution of crossveins in insect wings. J. Anat. 199:211-216.[Medline]

MCKIM, K. S. and A. HAYASHI-HAGIHARA, 1998  mei-W68 in Drosophila melanogaster encodes a Spo11 homolog: evidence that the mechanism for initiating meiotic recombination is conserved. Genes Dev. 12:2932-2942.[Abstract/Free Full Text]

MORGAN, T. H., 1910  Sex limited inheritance in Drosophila.. Science 32:120-122.[Free Full Text]

MORGAN, T. H., 1912  Complete linkage in the second chromosome of the male of Drosophila.. Science 36:719-720.

MORGAN, T. H., 1914  No crossing over in the male of Drosophila of genes in the second and third pairs of chromosomes. Biol. Bull. Woods Hole 26:195-204.

PAI, L. M., G. BARCELO, and T. SCHUPBACH, 2000  D-cbl, a negative regulator of the Egfr pathway, is required for dorsoventral patterning in Drosophila oogenesis. Cell 103:51-61.[Medline]

PRESTON, C. R. and W. R. ENGELS, 1996  P-element-induced male recombination and gene conversion in Drosophila.. Genetics 144:1611-1622.[Abstract]

PRESTON, C. R., J. A. SVED, and W. R. ENGELS, 1996  Flanking duplications and deletions associated with P-induced male recombination in Drosophila.. Genetics 144:1623-1638.[Abstract]

ROSS, J. J., O. SHIMMI, P. VILMOS, A. PETRYK, and H. KIM et al., 2001  Twisted gastulation is a conserved extracellular BMP antagonist. Nature 410:479-483.[Medline]

SLIZYNSKA, H., 1964  Location of crossveinless (cv, 13.7) on the salivary gland X-chromosome. Dros. Inf. Serv. 39:102.

WEIR, B. S., 1996 Genetic Data Analysis II: Methods for Discrete Population Genetic Data. Sinauer, Sunderland, MA.