Persistence time of a mutant allele, the expected number of generations before its elimination from the population, can be estimated as the ratio of the number of segregating mutations per individual over the mutation rate per generation. We screened two natural populations of Drosophila melanogaster for mutations causing clear-cut eye phenotypes and detected 25 mutant alleles, falling into 19 complementation groups, in 1164 haploid genomes, which implies 0.021 eye mutations/genome. The de novo haploid mutation rate for the same set of loci was estimated as 2 x 10^–4 in a 10-generation mutation-accumulation experiment. Thus, the average persistence time of all mutations causing clear-cut eye phenotypes is 100 generations (95% confidence interval: 61–219). This estimate shows that the strength of selection against phenotypically drastic alleles of nonessential loci is close to that against recessive lethals. In both cases, deleterious alleles are apparently eliminated by selection against heterozygous individuals, which show no visible phenotypic differences from wild type.

---

A substantial fraction of genetic variability of natural populations is due to unconditionally deleterious alleles, constantly produced by mutation and removed by selection (e.g., CROW and KIMURA 1970; KONDRASHOV and TURELLI 1992; SUNYAEV et al. 2001; RODRIGUEZ-RAMILO et al. 2004). At mutation-selection equilibrium, the per-individual number of segregating mutations is the product of the de novo mutation rate and the persistence time of the average mutation, P. Indeed, persistence time, defined as the average number of generations during which an allele survives in the population, is equal, in an infinite population, to the expected total number of carriers of the allele (CROW 1979, p. s172). Thus, the ratio of the per-individual number of segregating mutations over the mutation rate estimates P (CROW 1979; HOULE et al. 1996). If a mutation remains rare, the coefficient of selection hs against it is 1/P (CROW 1979).

In Drosophila, persistence time of recessive lethals is 50–100 generations, implying hs 0.01–0.02. Indeed, lethals appear with the rate 0.01–0.02/haploid genome/generation (see CROW and SIMMONS 1983; FRY et al. 1999), and a fly carries 0.5–1.5 heterozygous lethals per haploid genome (see LEWONTIN 1974; POWELL 1997), so that an organism carries 100 more lethals than what appears per generation. This classical analysis (CROW 1979; CROW and SIMMONS 1983) suggests that lethality is the only truly recessive manifestation of recessive lethals, since they must cause a substantial decline of fitness even when heterozygous.

However, recessive lethals represent only a fraction of deleterious spontaneous mutations, since most of such mutations are minor and do not have clear-cut phenotypic manifestations even when homozygous (TIMOFEEFF-RESSOVSKY 1935; MUKAI 1964; MUKAI et al. 1972). Indeed, only a minority of eukaryotic genes are essential, in the sense that loss of function is lethal, and many mutations of even essential genes do not disrupt the function completely (see KONDRASHOV et al. 2004). Still, MUKAI et al. (1972) and CROW (1979) claimed that in D. melanogaster, P and hs for minor deleterious mutations are similar to that of recessive lethals. In other words, recessive lethals differ from more common nonlethal mutations by how they affect homozygotes, but not heterozygotes.

However, estimates of P and hs for minor mutations depend on elaborate statistical analysis, which is a subject of ongoing controversy (see KEIGHTLEY 1994; KEIGHTLEY and EYRE-WALKER 1999). If free accumulation of minor mutations reduces fitness by much less than 1–2% per generation (the figure used by Crow), P for such mutations would substantially exceed that for recessive lethals. Thus, the extent of reduction of fitness caused by heterozygous loss-of-function mutations at nonessential genes remains unknown. It may be natural to assume that such mutations have measurable dominant effect on fitness, but the relevant data are currently absent.

We report data on persistence times of D. melanogaster mutations at loci where homozygous loss-of-function alleles produce clear-cut abnormal phenotypes, but are not lethal. Such "morbid" (KONDRASHOV et al. 2004) loci may be viewed as intermediate between those that can harbor recessive lethals and those where mutations could never cause obvious phenotypic effects. Specifically, we studied loci where drastic (mostly loss of function; YANG et al. 2001) alleles strongly affect color, texture, and/or shape of the eyes. We show that, similarly to recessive lethals, such mutations, although fully recessive with respect to their visible phenotypic effects, negatively affect fitness even when heterozygous.

The set of loci capable of producing eye phenotypes in D. melanogaster is well studied and functionally diverse (FREEMAN 1997; LLOYD et al. 1998; THOMAS and WASSARMAN 1999; WEASNER et al. 2005). FlyBase (FLYBASE CONSORTIUM 2003) lists between 150 and 160 such loci (depending on the trust given to "existence uncertain" loci), at a majority of which loss-of-function mutations are nonlethal (except 5 loci) and yet phenotypically overt. There are reasons to believe that a majority of such loci are already known: at least 75% of eye-color alleles found in natural populations affect the already known loci (APARISI and NAJERA 1990). The exact number of loci producing visible eye phenotypes does not matter for our purposes, as long as we measure the mutation rate and the standing crop of segregating alleles for the same set of loci.

Segregating eye-phenotype alleles were ascertained in the fall of 2004 for two wild populations, sampled at Countyside Winery (Blountville, TN) and on the campus of East Tennessee State University (30 km apart). F₂ inbred offspring of matings between unrelated parents from the same population were screened for eye phenotypes (TIMOFEEFF-RESSOVSKY and TIMOFEEFF-RESSOVSKY 1927). Females were obtained in the laboratory as daughters of wild-caught mothers (since most wild-caught females carry sperm), and males were wild caught. A total of 158 and 74 vials (i.e., 632 and 296 haploid genomes) were screened from the two populations. In addition, three mutations found in 59 sibships (236 haploid genomes) during the cleansing of the initial populations (see below) were added to the count of segregating mutant alleles. Since each vial screens mutations in 4 haploid genomes, the frequency of mutant alleles was estimated by dividing the number of observed alleles by 4N_v, where 4N_v is the number of processed vials. Confidence intervals of this frequency were estimated assuming that these frequencies were obtained by sampling a binomial population (ZAR 1999, p. 527).

Mutation rate has been estimated in a mutation-accumulation (MA) experiment, using offspring of wild-caught flies collected in the fall of 2002 at Countyside Winery (Blountville, TN). The founders of the MA populations were chosen by screening inbred F₂ offspring of wild-caught flies. Brothers and virgin sisters of the parents of the inbred offspring were retained to be used as outbred individuals to start the populations in case no segregating mutations were found among the inbred offspring. At least 15 (typically 19–23) replicate brother-sister matings were set for each pair of original wild-caught flies and in each vial at least 50 inbred offspring were screened.

If any of the parents contained a mutant allele, each of the replicate brother-sister matings had a 25% probability to contain a mutant phenotype and the frequency of this phenotype among the inbred offspring was, again, 25%. Assuming that we never missed a mutant phenotype in a vial where it occurs with the frequency of 25%, we overlooked a mutant present in the founder flies with the probability of 0.75^N, where N is the number of replicate brother-sister matings. In our cleansing effort, for each pair of founders we overlooked a mutation present in one of them with the probability between 0.0013 (23 replicates) and 0.0133 (15 replicates). Each of three replicate MA populations was started with seven pairs of unrelated founders. The probability of overlooking at least one mutant allele in these seven pairs combined is approximately the sum of probabilities of overlooking it in each of them. In our three populations such combined probabilities were 0.041, 0.051, and 0.055. Since the frequency of segregating eye phenotype alleles in the natural populations from which the founders were taken is low (see RESULTS), the chances of contamination of the MA populations by segregating alleles are negligible.

Experimental populations were maintained in 60-liter plastic boxes (Rubbermaid) with two mesh-covered windows cut in the opposite walls (3000-cm² total window area) and a mesh sleeve attached to the opening in the lid. Each generation was started by placing 12 plastic trays, each containing 250 ml of standard fly food medium, into each box and allowing the flies to lay eggs for 6 hr. After the eggs were laid, flies were partly immobilized by placing the boxes into a 4° chamber for 10 min, the trays were then removed and placed into a fresh box, and the flies were collected and weighed to estimate the population size. A hand-held compressed air blower (3M) was useful in removing the flies from the surface of the food and sides and bottoms of the trays. This procedure allowed us to maintain relatively stable, large populations: harmonic mean population size in the three replicate populations was 18,000, 21,000, and 17,000 flies.

Flies were sampled from generation 10. Offspring from each individual mating were mass mated and their combined inbred offspring screened for visible phenotypes, which corresponds to sampling 4 haploid genomes. All visible phenotypes were recorded and preserved, but only eye phenotypes counted toward the mutation rate estimate. Each vial was screened twice—within 24 hr from the eclosion of the first flies and 2–5 days later to assure detection of age-sensitive phenotypes. As long as selection against mutant phenotypes is weak, the expected frequency of mutant phenotypes in a vial in which at least 1 of 4 grandparental haploid genomes contained a mutation is 1/16, and therefore we overlooked a mutant with the probability of (1–1/16)^K, where K is the number of inbred offspring screen in each vial. At least 100 offspring were screened in each vial (typically well over 150), which results in missing a mutation with the probability of not greater than 0.0015. In other words, the probability of missing a mutant if it was sampled is negligible. This minor source of systematic error affects the mean but not the variance and therefore was not included in the calculation of confidence intervals. The total of 1401 vials representing 5604 haploid genomes was screened. Mutation rate per generation was estimated by dividing the number of observed mutant alleles by 4GN_v, where G is the number of generations in MA experiment (G = 10) and N_v is the number of vials screened. Just as in the population screen, confidence intervals around this estimate were obtained assuming sampling from a binomial population.

This approach is not sensitive to the bias introduced by clusters of premeiotic mutations (THOMPSON et al. 1998): repeatedly found mutations, even if possibly resulting from a premeiotic event, should all be counted toward the estimate of postmeiotic mutation rate per generation as long as the number of flies sampled is a small fraction of the effective population size (YANG et al. 2001).

Many dark-eye phenotypes are not clear cut and often appear to represent a continuum of shades and hues of reddish brown. To distinguish phenotypes caused by Mendelian alleles from individual phenotypic variability or alleles with low penetrance, all "suspect" eye phenotypes were bred and recovered in F₂, eliminating the possibility of false-positive detection. We ascertained suspect phenotypes liberally, and only approximately one-third of them turned out to be due to simple, Mendelian alleles. Thus, it is unlikely that many such alleles were overlooked. Since the same procedure was applied to both wild populations and MA screens, our estimates of P and hs should not be affected by possible rare errors. It should also be noted that since the two generations of inbreeding were reared under identical low-density conditions in both the MA experiment and the population screen, it is highly unlikely that any differences in nutrition or population density between the natural population and the MA cages could affect our ability to detect mutants in the two experiments.

Mutants found in the MA experiment and wild population screens were complemented to each other and to a set of phenotypically similar stocks. Mutants with dark-eye phenotypes were complemented against car, p, ry, or, bur, bw, ca, cho, mah, pr, rs, se, g, r, bo, ras, rb, dke, cl, pd, rud, and mal (22 stocks), mutants with bright- or diluted-eye phenotype against w, cn, st, v, z, kar, bri, cd, ltd, and lt (10 stocks), mutants with small and kidney-shaped eyes against as, k, hh, eyg, ey, eya, so, optix, and gv (9 stocks), and the rough-eye mutant against rh, rub, ro, gl, sina, and sv (6 stocks).

Persistence time was estimated as the ratio of the total frequency of segregating mutations to the total mutation rate per generation. Confidence interval for persistence time was estimated using quadratic sum of widths of upper and lower confidence intervals for mutation rate and frequency of segregating alleles, since errors in estimating these parameters are independent. To be conservative, the upper-bound confidence limit was estimated from the upper bound of the numerator and the lower bound of the denominator and the lower-bound confidence limit was estimated vice versa.

Mutants isolated in the population screens are listed in Table 1. The total number of independently observed mutants was 25, falling into 19 complementation groups, which corresponds to 0.0215 segregating mutations per haploid genome (95% C.I.: 0.015–0.030). Among the 19 complementation groups there were 14 singletons, 4 duplicates, and 1 triplicate. Of these 19 complementation groups, 13 had dark-eye phenotype (all autosomal), 3 had bright-eye phenotype (2 autosomal, 1 sex linked), and 3 had small-eye phenotype (all autosomal). A total of 4 of 13 dark-eye loci, all 3 bright-eye loci, and none of the 3 small-eye loci could be identified by complementation with stocks available from Bloomington Stock Center.

View this table: In this window In a new window   TABLE 1 Mutants found in population screens

 
In the MA experiment 12 independent mutants were recovered from the total of 1401 vials (5604 haploid genomes screened; Table 2). Of these 12 mutants, 9 had color phenotypes and all complemented each other. A total of 3 other mutants had a frontally notched or kidney-shaped eye and were found to be alleles of the same (yet unidentified) locus. Three of these 12 mutants were sex linked and 2 of these 3 were the only ones that could be identified by complementation with known stocks. None of the mutations found in the MA experiment failed to complement any of the mutations recovered from the populations.

View this table: In this window In a new window   TABLE 2 Mutants found in MA experiment

 
In addition to these 12 mutants, in 36 instances, spread over all three replicate populations, an additional phenotype was recovered, featuring rough eye texture with irregular ommatidia not arranged in straight lines. A total of 10 pairs involving 5 such rough eye lines (3 from population 1 and 1 from each of the other two populations) were tested for allelelism and found to be noncomplementing. This extraordinary high occurrence of mutations at the same locus is puzzling. If these mutants were observed in just one population, it could have been interpreted as a contamination of the founding population by a segregating allele or as a cluster caused by a premeiotic event. However, occurrence in all three replicates suggests that these observations are multiple independent events representing a hotspot for mutation, perhaps caused by transposition activity not present in the natural population (cf. HOULE and NUZHDIN 2004). Therefore, they were not counted toward the estimate of background mutation rate. This rough phenotype has never been recovered from the wild population, either during the population cleansing or during the screen for segregating mutants. Thus, the combined estimate of the mutation rate at all loci causing visible eye phenotypes was 2.0 x 10^–4/generation (95% C.I.: 1.2 x 10^–4–3.5 x 10^–4) or, assuming 150 loci, 1.4 x 10^–6/locus/generation (95% C.I.: 8.3 x 10^–7–2.3 x 10^–6) or 0.020 (0.010–0.030)/genome.

These estimates of the mutation rate and the standing crop of segregating mutations imply that an average mutant allele causing visible eye phenotype survives in the population for 100 generations (95% C.I.: 31–158). The same estimates for autosomal genes lead only to persistence time of 128 (25–180) generations and for sex-linked genes only to a (admittedly unreliable, based on only four mutations) value of 16 (0–76) generations.

In two natural populations of D. melanogaster, we detected 0.0215 eye-phenotype alleles per haploid genome, while the de novo mutation rate to such alleles, presumably at the same set of loci, was estimated as 2.0 x 10^–4/haploid genome. Thus, the average persistence time P of a mutant eye-phenotype allele is close to 100 generations. Most of such alleles are loss of function (YANG et al. 2001) and all are, with respect to their visible phenotype, fully recessive. Thus, our estimate of P for a set nonessential eye-color loci is only slightly, if at all, above the classic estimates of P for essential loci that harbor recessive lethals (50–100 generations, CROW 1979; CROW and SIMMONS 1983), and in both cases the harmonic mean coefficient of selection against heterozygotes, hs, is 1%.

Eye-color loci can be regarded as morbid (KONDRASHOV et al. 2004) since their mutations can cause, when homozygous, overtly abnormal, but not lethal phenotypes. In contrast, a majority of nonessential loci are probably also nonmorbid. However, estimates of P and hs for loci that are both nonessential and nonmorbid remain controversial, since individual mutations cannot be recognized. Our figure for P appears to be the first reliable estimate of persistence time for a subset of nonessential loci in Drosophila. To our knowledge, the only other estimate of the impact on fitness of loss-of-function alleles of nonessential loci is in good agreement with our data. THATCHER et al. (1998) observed that fitness of yeast with a nonessential knockout is down by 1–2%. These estimates are in remarkable agreement with the estimates of dominant deleterious effect of spontaneous mutations based on MA studies. LYNCH et al. (1999) reviewed several such studies in Drosophila and other organisms and concluded that an average de novo mutation decreases fitness by 2% in heterozygous state. Measurements of parameters of mutation in natural populations also often imply that a majority of mutations affecting fitness have substantial deleterious effect in heterozygous state (e.g., DENG and LYNCH 1997). Perhaps this agreement between estimates for nonessential genes and all genes affecting fitness reflects the fact that a large proportion of genes are nonessential, but do affect fitness.

The frequencies of eye phenotypes we observed in our populations are lower than those observed by APARISI and NAJERA (1990), who looked at a similar set of loci. Their average across six populations was two times higher than our estimate (60 mutant alleles in 1524 chromosomes). This difference, however, disappears if one locus, sf, with unusually high mutant allele frequencies (one-third of all found mutant alleles) and one of the populations (cellar) with unusually high frequency of mutants (possibly due to a special selection regime; NAJERA and MENSUA 1988) are removed from their sample.

Our data are in excellent agreement with those of APARISI and NAJERA (1990) with respect to the degree of saturation. With Poisson distribution with equal frequencies assumed and the size of unobserved null class of the Poisson distribution from the frequencies of singletons and duplicates estimated, both studies result in the estimate of 45 segregating loci. Of course, this is a lower-bound estimate, since mutant alleles at different loci certainly are segregating at different frequencies.

Mutation rate per locus measured in our MA experiment (1.4 x 10^–6) is somewhat lower than many locus-specific estimates, which typically range between 10^–6 and 5 x 10^–5. This may simply reflect a broad confidence interval around our estimate or an overestimation of the number of loci we have screened. It may also reflect the fact that locus-specific estimates are often based on a biased set of loci (highly mutable loci are the easiest to study) and may include hotspots similar to the one we excluded in our analysis. Note that the large uncertainty introduced by unknown number of loci does not bias out persistence time estimate. On the other hand, our figure for de novo mutation rate and, thus, for P, may be too low, since there must be some selection against heterozygous eye-phenotype alleles in our cage populations. However, we estimate that hs acting against such mutations was 0.01. The impact of such selection in the course of only 10 generations cannot be large.

Thus, we do not see any large difference between heterozygous effects of mutations at essential and nonessential, although morbid, loci in D. melanogaster, which is consistent with the conclusions of CROW (1979). Of course, the possibility that heterozygous effects of mutations at loci that are both nonessential and nonmorbid are milder still cannot be ruled out. Yet, one conclusion is clear: selection against heterozygous loss-of-function mutations at eye loci is not significantly different from that against heterozygous recessive lethals.

If these results are typical for all nonessential but morbid loci, they should be considered in the context of the rate of evolution in essential and nonessential genes. Interspecies genome comparison suggests that essential (morbid) genes evolve 40% more slowly than most dispensible (nonmorbid) genes (HIRSH and FRASER 2001; KONDRASHOV et al. 2004). However, when essential (i.e., loss-of-function lethal) genes are compared to genes in which loss-of-function mutations have moderate to strong effects, but are not lethal (HURST and SMITH 1999; YANG et al. 2003), little difference in the rates of evolution can be detected. This is exactly the prediction suggested by the results of this study, although such comparisons must be done with care, because differences in rate of evolution reflects differences in the rate of fixation of beneficial, neutral, and mildly deleterious mutations, while mutations observed in this study are too harmful to ever be fixed.

We are grateful to Jeff Stirman, Gloria Simmons, and Aubri Shaver for laboratory assistance, to Justin Kumar and his lab for conducting a number of complementation tests, and to David Houle and two anonymous reviewers for helpful suggestions and criticisms. L.Y. and A.C. were supported by East Tennessee State University's RDC and HHMI grants.

APARISI, M. L., and C. NAJERA, 1990 Genetic location and biochemical characterization of eye-colour mutants from natural populations of Drosophila melanogaster. Genome 33: 203–208.[Medline]

CROW, J. F., 1979 Minor viability mutants in Drosophila. Genetics 92: s165–s172.[Medline]

CROW, J. F., and M. KIMURA, 1970 An Introduction to Population Genetics Theory. Burgess Publishing, Minneapolis.

CROW, J. F., and M. J. SIMMONS, 1983 The mutation load in Drosophila, pp. 1–35 in The Genetics and Biology of Drosophila, Vol. 3c, edited by M. ASHBURNER, H. L. CARSON and J. N. THOMPSON, JR. Academic Press, London.

DENG, H. W., and M. LYNCH, 1997 Inbreeding depression and inferred deleterious-mutation parameters in Daphnia. Genetics 147: 147–155.[Abstract]

FLYBASE CONSORTIUM, 2003 The FlyBase database of the Drosophila genome projects and community literature. Nucleic Acids Res. 31: 172–175.[Abstract/Free Full Text]

FREEMAN, M., 1997 Cell determination strategies in the Drosophila eye. Development 124: 261–270.[Abstract]

FRY, J. D., P. D. KEIGHTLEY, S. L. HEINSOHN and S. V. NUZHDIN, 1999 New estimates of the rates and effects of mildly deleterious mutation in Drosophila melanogaster. Proc. Natl. Acad. Sci. USA. 96: 574–579.[Abstract/Free Full Text]

HIRSH, A. E., and H. B. FRASER, 2001 Protein dispensability and rate of evolution. Nature 411: 1046–1049.[CrossRef][Medline]

HURST, L. D., and N. G. SMITH, 1999 Do essential genes evolve slowly? Curr. Biol. 9: 747–750.[CrossRef][Medline]

HOULE, D., B. MORIKAWA and M. LYNCH, 1996 Comparing mutational variabilities. Genetics 143: 1467–1483.[Abstract]

HOULE, D., and S. V. NUZHDIN, 2004 Mutation accumulation and the effect of copia insertions in Drosophila melanogaster. Genet. Res. 83: 7–18.[CrossRef][Medline]

KEIGHTLEY, P. D., 1994 The distribution of mutation effects on viability in Drosophila melanogaster. Genetics 138: 1315–1322.[Abstract]

KEIGHTLEY, P. D., and A. EYRE-WALKER, 1999 Terumi Mukai and the riddle of deleterious mutation rates. Genetics 153: 515–523.[Abstract/Free Full Text]

KONDRASHOV, A. S., and M. TURELLI, 1992 Deleterious mutations, apparent stabilizing selection and the maintenance of quantitative variation. Genetics 132: 603–618.[Abstract]

KONDRASHOV, F. A., A. Y. OGURTSOV and A. S. KONDRASHOV, 2004 Bioinformatical assay of human gene morbidity. Nucleic Acids Res. 32: 1731–1737.[Abstract/Free Full Text]

LEWONTIN, R. C., 1974 The Genetic Basis of Evolutionary Change. Columbia University Press, New York.

LLOYD, V., M. RAMASWAMI and H. KRAMER, 1998 Not just pretty eyes: Drosophila eye-colour mutations and lysosomal delivery. Trends Cell Biol. 8: 257–259.[CrossRef][Medline]

LYNCH, M., J. BLANCHARD, D. HOULE, T. KIBOTA, S. SCHULTZ et al., 1999 Perspective: spontaneous deleterious mutation. Evolution 53: 645–663.[CrossRef]

MUKAI, T., 1964 The genetic structure of natural populations of Drosophila melanogaster. I Spontaneous mutation rate of polygenes controlling viability. Genetics 50: 1–19.[Free Full Text]

MUKAI, T., S. I. CHIGUSA, L. E. METTLER and J. F. CROW, 1972 Mutation rate and dominance of genes affecting viability in Drosophila melanogaster. Genetics 72: 335–355.[Abstract/Free Full Text]

NAJERA, C., and J. L. MENSUA, 1988 Stable polymorphism for mutant eye colour genes in populations of Drosophila melanogaster in two different media. Genetica 77: 123–131.[Medline]

POWELL, J. R., 1997 Progress and Prospects in Evolutionary Biology: The Drosophila Model. Oxford University Press, New York.

RODRIGUEZ-RAMILO, S. T., A. PEREZ-FIGUEROA, B. FERNANDEZ, J. FERNANDEZ and A. CABALLERO, 2004 Mutation-selection balance accounting for genetic variation for viability in Drosophila melanogaster as deduced from an inbreeding and artificial selection experiment. J. Evol. Biol. 17: 528–541.[CrossRef][Medline]

SUNYAEV, S., V. RAMENSKY, I. KOCH, W. LATHE, A. S. KONDRASHOV et al., 2001 Prediction of deleterious human alleles. Hum. Mol. Genet. 10: 591–597.[Abstract/Free Full Text]

THATCHER, J. W., J. M. SHAW and W. J. Dickinson, 1998 Marginal fitness contributions of nonessential genes in yeast. Proc. Natl. Acad. Sci. USA. 95: 253–257.[Abstract/Free Full Text]

THOMAS, B. J., and D. A. WASSARMAN, 1999 A fly's eye view of biology. Trends Genet. 15: 184–190.[CrossRef][Medline]

THOMPSON, JR., J. N., R. C. WOODRUFF and H. HUAI, 1998 Mutation rate: a simple concept has become complex. Environ. Mol. Mutagen. 32: 292–300.[CrossRef][Medline]

TIMOFEEFF-RESSOVSKY, N. W., 1935 Auslosung von Vitlitatsmutationen durch Rontgenbestrahlung bei Drosophila melanogaster. Nachr. Gess Wiss. Gottingen Biol. N. F. 1: 163–180.

TIMOFEEFF-RESSOVSKY, H. A., and N. W. TIMOFEEFF-RESSOVSKY, 1927 Genetische Analyse einer freilebenden Drosophila melanogaster populations. Arch. Entwicklungsmech. Org. 109: 70–109.[CrossRef]

WEASNER, B., L. ANDERSON and J. P. KUMAR, 2005 Eye specification in Drosophila. Proc. Indian Natl. Sci. Acad. (in press).

YANG, H. P., A. Y. TANIKAWA and A. S. KONDRASHOV, 2001 Molecular nature of 11 spontaneous de novo mutations in Drosophila melanogaster. Genetics 157: 1285–1292.[Abstract/Free Full Text]

YANG, J., Z. GU and W. H. LI, 2003 Rate of protein evolution versus fitness effect of gene deletion. Mol. Biol. Evol. 20: 772–774.[Abstract/Free Full Text]

ZAR, J. H., 1999 Biostatistical Analysis. Prentice-Hall, Upper Saddle River, NJ.

Communicating editor: D. HOULE

International Access Link

Online ISSN: 1943-2361  Print ISSN: 0016-6731
Copyright 2009 by the Genetics Society of America
phone: 412-268-1812   fax: 412-268-1813   email: genetics-gsa{at}andrew.cmu.edu