Complementation Analysis How Many Genes are Involved

Complementation analysis is used to determine whether two independent mutations are alterations in the same gene; that is, they are alleles, or are alterations in different genes. In essence, a complementation analysis is a functional test used to define a gene. If a researcher has isolated a number of mutants with a similar phenotype, the next question asked is: 'How many genes have I identified?'. If there are 10 mutant strains, are they each in different genes, does each mutant carry a different mutation (allele) in the same gene, or something in between such as two genes one with six alleles and the other with four alleles? Complementation analysis will help answer this question.

Seymour Benzer's study of the rll locus of phage T4 is a most elegant example of the power of complementation analysis (Benzer, 1955). Benzer had several hundred mutations that gave the same phenotype, large plaques on one host strain of E. coli and no plaques on another host strain, and mapped to the same region of the T4 chromosome. Using the host strain in which rll mutants formed no plaques, he found that when host cells were coinfected with different mutant pairs some pairs produced a normal phage burst while others did not. Those that produced a normal burst he concluded were in different functional genetic units that he called cistrons, a term that is synonymous with gene. In contrast, those pairs of mutants that rarely or never produced a normal burst he concluded were in the same cistron. The rare productive infections Benzer proposed resulted from recombination between the different mutations in the same cistron thereby creating a wild-type genotype in a few coinfected cells. Using this method, he placed all of his rll mutations into two cistrons that he called rllA and rllB. Moreover, he made a detailed genetic map based on recombination frequency between the different mutations (a fine structure map) that also indicated the frequency at which a mutation was isolated at that position. In this way he demonstrated that genes are not indivisible units but consist of many mutable sites that can recombine. Geneticists working with other organisms soon followed Benzer's lead and adapted complementation analysis to their systems.

To carry out a complementation analysis, both mutant genes must be expressed in the same cell so that their gene products are synthesized in the same cytoplasm and can functionally interact. Only loss of function (recessive) mutations can be used for a complementation analysis. The theory behind the complementation analysis is simple. If both mutations are loss of function alterations of the same gene, then the diploid cell carrying these two mutant genes will not contain a functional allele and will have the mutant phenotype. If both mutations are loss of function mutations but in different genes, then the diploid cell will have one mutant allele and one functional wild-type allele of each gene and, since the mutant alleles are loss of function alleles, the diploid will have the wild-type phenotype.

To express both mutant genes in the same cytoplasm a heterozygous diploid must be constructed. The way the researcher establishes the diploid state varies with the organism under study. In Saccharomyces, this is accomplished by mating a MATs strain containing mutation #1 to a MATa strain containing mutation #2. The a la diploid will be heterozygous for the mutant genes. The phenotype of the heterozygous diploid is then observed. If the diploid has a wild-type phenotype, then the mutations are said to complement and this is strong evidence that the mutations are in different genes. A geneticist might also say, 'The mutations are in different complementation groups'. If the diploid has a mutant phenotype, then the mutations do not complement and are said to be in the same complementation group. This is considered strong evidence that the mutations are alleles. The definition of complementation group is a set of noncomplementing mutations. The term complementation group is synonymous with gene.

Two conditions must be met before one can carry out a complementation analysis on a series of mutants. First, the mutant strain can only contain a single mutation compared with the parental strain. Particularly when mutagenesis had been used to obtain the mutants, it is possible that more than one DNA alteration was induced in an individual and these are involved in producing the mutant phenotype. Therefore, each mutant strain must be tested to demonstrate whether one or more genetic alterations are required to produce the mutant phenotype. To test this the mutant strain is crossed to a wild-type strain and tetrad analysis of the heterozygous diploid is carried out. If only a single alteration is required, then only tetrads with two mutant spores and two wild-type spores will be produced, as shown in Cross 1 in Chapter 1. But if two or more alterations are present, tetratype and nonparental ditype tetrads will be produced, as shown in Cross 3 in Chapter 1. What would be the phenotypes of the spores of a tetratype tetrad if the mutant strain contained two altered genes and both alterations were required to produce the mutant phenotype? What would be the phenotypes of the spores of a tetratype tetrad if the mutant strain contained two altered genes and either mutation alone were sufficient to produce the mutant phenotype?

Of course to carry out a cross between the mutant and wild-type strains the strains must be of opposite mating type and should carry different nutritional mutations to facilitate the selection of diploids. But in other respects the two strains should ideally be isogenic except for any alterations required to produce the mutant phenotype. Usually, before undertaking a mutant hunt, the geneticist will construct an appropriate pair of isogenic (or congenic) haploid strains to be used as parental strains. The mutants isolated in one strain can then be mated to the parental strain of the opposite mating type to determine the number of mutant genes involved.

As described above, the second requirement for a complementation analysis is that the mutations be loss of function alleles. In other words, only mutations that are recessive to the wild-type allele can be used. So, as a second step in the genetic analysis of mutants, mutant strains carrying a single mutant gene are crossed to a parental strain carrying the wild-type allele. If the mutant carries a recessive loss of function mutation, then the heterozygous diploid (GENllgenl-34) will have the wild-type phenotype. This mutant allele can then be used for complementation analysis.

Cross 4 shows a complementation test for two mutant strains. Preliminary genetic analysis has shown that each strain contains only a single mutant gene and that the mutant allele is recessive.

Cross 4:

Diploid phenotype:

Mutant strain 5 Mutant

Mutant strain 14

The result shown in Cross 4 indicates that the mutation in strain 5 and the mutation in strain 14 do not complement and thus are mutations in the same gene. If we call the gene GEN J, then these mutations are alleles and one could now name them genl-5 and genl-14. This cross could be depicted as shown below.

Cross 4: genl-5 x genl-14 (genotypes of parental strains)

(mutant) (mutant) (phenotype of parental strains)

Diploid: genl-5

genl-14 (mutant)

(genotype of diploid)

(phenotype of diploid)

As a second test of whether or not the mutations are alleles, the researcher can determine the segregation pattern of the alleles in the meiotic products of the diploid. If the two mutations are in the same gene, then recombination between the mutations will be relatively rare because they map so close to one another. Therefore, 100% of the time (or close to it) the two mutant genes will segregate to different spores producing a tetrad with four mutant spores (two mutant #5 spores and two mutant #14 spores). This situation is depicted in Cross 2 of Chapter 1).

Cross 5 shows another complementation test between mutant strain 5, which carries the mutation genl-5, and another mutant strain. Mutant strain 4 contains only a single mutant gene and the mutant allele is recessive.

Cross 5:

Diploid phenotype:

Mutant strain 5 Wild-type

Mutant strain 4

The result shown in Cross 5 indicates that the mutation in strain 5 and the mutation in strain 4 complement and thus are mutations in different genes. We can then say that a different gene, GEN2, is mutant in strain 4. This cross could be depicted as shown below.

Cross 5: genl-5 GEN2 x GEN1 gen2-4 (genotypes of parental strains) (mutant) (mutant) (phenotype of parental strains)

Diploid: genl-5 gen2-4

GENI GEN2 (wild-type)

(genotype of diploid) (phenotype of diploid)

If GEN1 and GEN2 are not linked, then the mutant genes will recombine producing recombinant meiotic products with the wild-type (GEN1 GEN2) and double mutant (genl-5 gen2-4) genotype. This is exhibited by the presence of tetratype and nonparental ditype tetrads when this diploid is subjected to tetrad analysis (see Cross 3 in Chapter 1). The frequency of each type of tetrad will depend on the frequency of recombination. If the two genes are completely unlinked, that is 50% recombination, the frequency of PD: TT: NPD tetrads will be 1:4:1. If there is any linkage, then the frequency of recombination is less than 50% and the relative number of PD tetrads will increase to greater than the expected 1/6 of the total number of tetrads analyzed. Ultimately, for crosses between two alleles, the number of PD tetrads will closely approach 100%, as is shown in Cross 1. One can calculate the map distance between two mutations (the frequency recombination multiplied by 100) using the following formula, which is correct for map distances up to 35 cM (Sherman & Wakem, 1991):

'6 x # NPD tetrads + # TT tetrads" total # of tetrads

The combination of these two methods, complementation analysis and tetrad analysis, should clearly indicate whether one is dealing with mutations in one or more genes. Either method alone is not as powerful, and therefore researchers do both tests. For example, if there are mutations in two very tightly linked genes, the mutations will complement, but recombination will be rare and most, if not all the tetrads will be PD. Such results would strongly suggest that one is dealing with mutations in two closely linked genes.

In a complete complementation analysis, all the mutants are crossed to all of the other mutants. Often as a complementation group containing several mutant alleles is identified, one allele will be chosen as the representative of that complementation group and only this allele will be crossed to the other mutants. At the end of this process, all the mutants isolated in a particular mutant selection/screen will be placed into complementation groups, i.e. genes. The researcher will have made a good start at determining the number of genes involved in the process of interest. If only a few genes have been identified with several mutant alleles, then the researcher will have some degree of confidence that the analysis saturated the genome and that new genes are not likely to be identified by the same selection/screening method. If many genes have been identified, several with only one mutant allele each, then it is likely that new genes will be identified if the same selection/screen is repeated.

There are some special situations in which straightforward interpretation of a complementation test is misleading and the geneticist must be on the alert for such possibilities. Infrequently, mutant alleles of the same gene are able to complement, and produce a heterozygous diploid with a wild-type like phenotype. This is referred to as intragenic complementation. Intragenic complementation can occur if the encoded polypeptide forms a multiple subunit protein composed of like subunits, such as a homodimer, or if it encodes a single polypeptide that carries out several distinct functions. In the case of the homomultimeric protein, mutant subunits encoded by different mutant genes associate with one another in the multimeric protein and are able to accommodate each other's mutant alteration in the mixed w 100

map distance in cM =

multimer. When this happens the mixed-mutant multimer complex has some functional activity, although it may not be completely normal.

Another mechanism of intragenic complementation is possible if the protein product of the gene has several distinct functions, such as two different enzyme activities. In this situation it is possible to obtain mutations that affect one of these activities while leaving the other function intact. In a heterozygous diploid, cells carrying two different mutations, each one affecting only one of the two functions, proteins capable of carrying out both enzyme activities will be produced, albeit in different molecules, and the cell should have the wild-type phenotype.

In contrast to intragenic complementation where mutations in the same gene complement, in a few instances mutations in different genes which are expected to complement do not. This phenomenon is referred to as nonallelic noncomplementa-tion. One explanation for this noncomplementation is that the two genes encode subunits of a heteromultimeric protein, and that the presence of a mutant alteration in either subunit destroys all function of the multimeric protein. Sort of, 'one bad apple spoils the whole barrel'.

Careful and thorough genetic analysis involving both complementation tests and genetic mapping of several mutant alleles is necessary to avoid the pitfalls of these potentially misleading situations.


Benzer, S. (1995) Fine structure of a genetic region in bacteriophage. Proc. Natl Acad. Sci.

Sherman, F. & P. Waken (1991) Mapping yeast genes. Methods Enzymol. 194: 38-57.

Genetic Techniques for Biological Research Corinne A. Michels Copyright © 2002 John Wiley & Sons, Ltd ISBNs: 0-471-89921-6 (Hardback); 0-470-84662-3 (Electronic)

6 Epistasis Analysis


The Random House Dictionary of the English Language—Unabridged Edition (1966) defines epistasis as a genetic term describing the 'interaction between nonallelic genes in which one combination of such genes has a dominant effect over other combinations'. The key word to remember is 'nonallelic', i.e. different genes. Epistasis (from the Greek meaning 'stand above') is the masking of the phenotype of a mutation in one gene by the phenotype of a mutation in another gene (Huang & Sternberg, 1995). One gene is said to be epistatic to another when the double mutant strain exhibits the phenotype of that mutant gene. This is in clear contrast to the terms dominant and recessive, which describe the relationship between different alleles of the same gene. It is very important not to confuse these concepts.

Epistasis analysis is used to determine if genes with related mutant phenotypes act in the same or different pathways, and, if in the same pathway, to place them in a linear order relative to one another based on the step in the pathway controlled by that gene. In other words, one uses epistasis analysis to construct an order-of-function map that reflects the sequence of events in a pathway controlled by several genes.

The use of epistasis analysis for the study of complex pathways was suggested more or less simultaneously by two independent research groups working in different fields. Jarvik & Botstein (1973) reported the isolation of temperature-sensitive and cold-sensitive mutations that block phage P22 morphogenesis, i.e. assembly of the phage particle. They used a combination of double mutant studies and reciprocal temperature shifts (made possible by their use on conditional ts and cs mutants) to determine the order of events in phage assembly. Their work demonstrated that head and tail assembly were independent processes but that both were dependent on phage DNA replication. Hereford & Hartwell (1974) used epistasis analysis to order events in the Saccharomyces cell division cycle. They used temperature-sensitive mutants that blocked the cell cycle at morphologically distinguishable points. The execution points of these genes were ordered relative to the block produced by the cell cycle inhibitor a-factor by temperature-shift experiments and relative to each other by double mutant studies.

To determine the epistatic relationship between two genes, mutations in these genes must have distinguishable phenotypes. Epistasis analysis is undertaken only after the initial steps of genetic analysis. Mutations are isolated and placed into complementation groups. Then, representative alleles are selected for a detailed characterization of phenotype so as to reveal subtle differences in phenotype not obvious from initial characterization. This could be a complete morphological analysis, such as in secretory pathway mutants or cell division cycle mutants, or intensive biochemical characterization, such as for DNA replication mutants or mutants affecting metabolic pathways. Occasionally, different alleles exhibit somewhat different phenotypes. The researcher can now capitalize on these phenotypic differences in the epistasis analysis.

For the purposes of epistasis analysis, there are two types of pathway in living systems, substrate-dependent pathways and switch regulatory pathways (Huang & Sternberg, 1995). A substrate-dependent pathway consists of an obligate series of steps or reactions that are required to produce a final outcome. The outcome can be as simple as the synthesis of a nutrient such as an amino acid or a macromolecule, or can be as complex as the formation of a ribosome. The substrate-dependent pathway can be thought of as a progression of events or even as a river flowing downstream with separate tributaries joining at different points and finally flowing into the lake. Another view is as a series of positive reactions each dependent on a source of substrate and a functional gene product for the successful completion of each step in the pathway. Moreover, the product of the more upstream step is used as the substrate of the downstream step. If there is no substrate available or if any one of the enzymes is missing or inactive, then the pathway will be blocked at that step. A production line at a factory would be considered to be this type of pathway. It is dependent on the input of parts (substrates) and workers (gene products) to assemble these parts.

A switch regulatory pathway consists of a series of genes or gene products that alternate between two states, 'on' and 'off'. The components of this pathway are usually acting directly on each other as opposed to on substrates, as occurs in a substrate-dependent pathway. The activity of a switch regulatory pathway is regulated by an upstream signal that stimulates the pathway and produces the downstream response. Environmental changes, cell-cell interactions, zygote formation, and mitogenic signals are only a few of the signals that can act as initiators of a switch regulatory pathway. The downstream response can be altered gene expression, cell division, or the initiation of a developmental process such as pattern formation.

Mutations in the genes encoding components of a switch regulatory pathway can lock the component into a permanently 'on' or permanently 'off' state. This has the effect of separating the downstream response from the initiating signal. Mutations that allow the response to be produced even in the absence of a stimulatory signal or despite the presence of an inhibitory signal are referred to as constitutive mutations. The isolation of constitutive mutations is a strong indicator that one is dealing with a switch regulatory pathway.

In a switch regulatory pathway the action of a particular component, or regulatory factor, can be either positive or negative. The function of a positive regulatory factor is to activate the next component in the pathway (its downstream component) when it is in its active form. A recessive (loss of function) mutation in a gene encoding a positive regulatory factor blocks the pathway. A dominant (gain of function) mutation in a gene encoding a positive regulatory factor produces a protein capable of functioning constitutively even in the absence of upstream activation of the pathway. A negative regulatory factor in the activated state inactivates the next downstream component in the pathway. Therefore, a recessive, loss of function mutation in a gene encoding a negative regulatory factor allows the next step in the pathway to be constitutively active. A dominant, gain of function mutation in a gene encoding a negative regulatory factor produces a gene product capable of constitutively inhibiting the pathway even in the absence of the signal. By determining whether the mutation is dominant or recessive, constitutive or blocks the response, the geneticist can decide whether the gene product is a negative or positive regulatory factor.

Determining whether the pathway under investigation is a substrate-dependent or switch regulatory pathway is not a simple task. Nonetheless, as will be seen below, it is important because the interpretation of the results of double mutant analysis depends on the type of pathway. The characterization of mutant alleles of pathway genes can provide some clues. As discussed above, if constitutive mutants are obtained in any of the pathway genes, then one can conclude that the pathway is a switch regulatory pathway. More often than not, a switch regulatory pathway controls more than one downstream response. As a result mutations in a switch regulatory pathway affect a number of phenotypic traits, such as the expression of several genes, and are said to be pleiotropic. The identification of pleiotropic mutants is suggestive of a switch regulatory pathway. If no constitutive mutations in pathway components have been identified, one can proceed with an epistasis analysis under the assumption that one is dealing with a substrate-dependent pathway. For complex processes this is unlikely to be the case. More likely, the initial characterization of mutations has not uncovered all the genes in the pathway or isolated a sufficiently varied array of mutant alleles. As the genetic analysis of the pathway proceeds new genes and/or new alleles will be identified and the true character of the pathway will be revealed in full detail. This will become clearer as examples from the literature are discussed.


Let us say that GEN1 and GEN2 are related because mutations in both genes decrease the production of Z. Mutations in GEN1 give phenotype A, and mutations in GEN2 give phenotype B. Only recessive loss of function alleles of GEN1 and GEN2 have been isolated. No constitutive alleles of GEN1 or GEN2 have been identified and mutations in these genes alter Z production but appear not to affect other phenotypes. We assume that this is a substrate-dependent pathway and proceed with the epistasis analysis.

The four mechanisms of genetic interaction between GEN1 and GEN2 are given in Table 6.1 (based on Hereford & Hartwell, 1974) and the resulting phenotype of the single or double mutation with regard to production of Z is indicated.

Models 1 and 2: Proteins Genlp and Gen2p participate in different steps of the same pathway and protein Genlp acts in a step that is upstream (Model 1) or downstream (Model 2) of the step catalyzed by protein Gen2p.

Model 3: Proteins Genlp and Gen2p are components of two independent and parallel pathways for Z production.

Model 4: Proteins Genlp and Gen2p act at the same step and in conjunction with one another.

To determine the epistatic relationship between these two genes, one constructs a strain that is mutant at both genes and observes the phenotype of the double mutant


Table 6.1 Epistasis analysis of a substrate-dependent pathway


Table 6.1 Epistasis analysis of a substrate-dependent pathway

Phenotype of geni single mutation

Phenotype of gen2 single mutation

Phentotype of genJ gen2 double mutation

Model 1





Model 2





Model 3





Model 4




=A =B or unique

strain. If the double mutant phenotype is A, then GEN1 is epistatic to GEN2 and GEN1 encodes the upstream component in the pathway. Alternately, if the double mutant phenotype is B, then GEN2 is epistatic to GEN1 and GEN2 encodes the upstream component. In summary, in a substrate-dependent pathway, if the double mutant exhibits a phenotype identical to one or the other mutant genes, then that gene is epistatic and encodes the more upstream component in the pathway.

If Model 3 or 4 describes the relationship, the results can be more difficult to interpret. The term 'unique' used in Table 6.1 indicates that the phenotype is different from either phenotype A or B. It can be qualitatively related to phenotypes A and B but quantitatively more extreme. For example, mutations in two different RAD genes might partially reduce the rate of recombination but to different extents, while the double mutant completely blocks all recombination. This type of interaction is called enhancement and will be discussed in detail in Chapter 9. Another classic example of two parallel pathways affecting a single trait comes from Drosophila. The reddish brown eye color results from a mixture of two pigments synthesized by parallel pathways. Mutations in one of these pathways that blocks red pigment production produces brown eyes while mutations in the alternate pathway that blocks brown pigment production produces red eyes. Flies defective for the production of both pigments, that is double mutants, have white eyes. White eyes is a unique phenotype and could not have been predicted from observing the phenotypes of the single mutants. The double mutant in Model 4 might exhibit a unique phenotype or could be phenotype A or B. What will distinguish Model 4 from Models 1 and 2 is that the phenotype of the double mutant is likely to vary with the alleles. Thus, epistasis analysis has limitations. The researcher will have to proceed to other methods, such as suppressor and enhancer analysis or coimmuno-precipitation, to support a proposed model.


In a switch regulatory pathway, the epistatic gene encodes the downstream component. Because this type of pathway involves both negative and positive components, it is not possible to come up with a simple table describing all possible results. Instead, several examples of hypothetical switch regulatory pathways will be presented for the reader to ponder. Then some sample results will be described and the reader can practice analytical skills. In the pathways given below, an arrowhead indicates that the action of the protein or signal is positive and a vertical line indicates that the action of the protein or signal is negative. For each pathway the reader should determine the phenotype of mutants in each protein component. The classes of mutations to consider are alleles that put the component in the permanently 'on' state and those that put it in the permanently 'off' state. The possible phenotypes are constitutive or no response produced in the presence of signal.

Regulatory pathway 1

Signal-Protein A-*■ Protein B-Protein C-► Response

Regulatory pathway 2

Signal-Protein X-1 Protein Y-1 Protein Z-Response

The experiments listed below present the results of epistasis analysis of mutations in genes GEN], GEN2, GEN3, and GEN4. Mutations in these genes affect the same process and may be in a common pathway. A number of alleles of each are available including constitutive alleles. Experiments 1-5 provide an example of an epistasis analysis of a switch regulatory pathway. Use these results as a practice exercise. Determine whether each gene encodes a positive or negative regulator and the order-of-function of the genes in the pathway. As a guide, the results of each experiment are followed by their interpretation. Regulatory pathway 3 synthesizes these conclusions into an order-of-function map of the pathway.

Experiment 1: A recessive mutation in GEN1 does not respond to the signal. A recessive mutation in GEN2 is constitutive. A strain carrying both mutations (gen] gen2) is constitutive. {Conclusions: Genl protein is a positive regulator. Gen2 protein is a negative regulator. GEN2 is epistatic to GEN] and the product of GEN2 acts downstream of the product of GEN].)

Experiment 2: A recessive mutation in GEN2 is constitutive. A recessive mutation in GEN4 blocks the response to the signal. A strain carrying both mutations (gen2 gen4) does not respond to the signal. (Conclusions: Gen4 protein is a positive regulator. GEN4 is epistatic to GEN2 and the product of GEN4 acts downstream of the product of GEN2.)

Experiment 3: A dominant constitutive allele of GEN1 is identified. A recessive mutation in GEN3 blocks the response to the signal. A strain carrying both mutations (GENl-c gen3) does not respond to the signal. (Conclusions: Genl-c protein is an activated form of the protein that is locked in the 'on' state. Gen3 protein is a positive regulator. GEN3 is epistatic to GEN1 indicating that the product of GEN3 acts downstream of the product of GEN1.)

Experiment 4: A dominant constitutive allele of GEN4 is isolated. A recessive mutation in G EN3 blocks the response to the signal. A strain carrying both mutations (gen3 GEN4-c) is unable to respond to the signal. (Conclusion: GEN3 is epistatic to GEN4 and the product of GEN3 acts downstream of the product of GEN4.)

Experiment 5: A dominant constitutive allele of GEN3 is isolated. A recessive mutation in GEN4 blocks the response to the signal. A strain carrying both mutations (GEN3-C gen4) does not respond to the signal. (Conclusion: GEN4 is epistatic to G EN3. This result taken together with the results of Experiment 4 suggests that the products of GEN3 and GEN4 could act at the same step.

Regulatory pathway 3


Genes encoding proteins that function in the same pathway or process will exhibit epistasis, as defined in the discussion presented above. Such genes are said to be members of an epistasis group. A well-studied example of an epistasis group is the RAD52 epistasis group (Pâques & Haber, 1999). Mutations in genes such as RAD50, RAD51, RAD52, RAD54, RAD55, RAD57, RAD59, SPOll, and M RE 11 exhibit defects in recombination and double-strand break repair. Construction of double mutant strains demonstrates epistatic relationships among the various members of this group of genes. Mutations in RAD3 or RAD6, which like RAD52 were originally isolated because of their increased sensitivity to X-ray radiation, do not exhibit epistasis with the RAD52 epistasis group or with each other. When paired with RAD52 or other members of the RAD52 epistasis group, the results are consistent with Model 3 in Table 6.1 and clearly indicate that RAD3 and RAD6 are in distinct pathways. Thus, despite certain similarities in phenotype, RAD52, RAD3, and RAD6 are members of different epistasis groups.


Avery, L. & S. Wasserman (1992) Ordering gene functions: the interpretation of epistasis in regulatory hierarchies. Trends Genet. 8: 312-316. Botstein, D. & R. Maurer (1982) Genetic approaches to the analysis of microbial development. Ann. Rev. Genet. 16: 61-83. Hereford, L.M. & L.H. Hartwell (1974) Sequential gene function in the initiation of

Saccharomyces cerevisiae DNA synthesis. J. Mol. Biol. 84: 445-461. Huang, L.S. & R.W. Sternberg (1995) Genetic dissection of developmental pathways.

Methods Cell Biol. 48: 97-122. Jarvik, J. & D. Botstein (1973) A genetic method for determining the order of events in a biological pathway. Proc. Natl Acad. Sci. USA 70: 2046-2050. Pâques, F. & J.E. Haber (1999) Multiple pathways of recombination induced by double-stranded breaks in Saccharomyces cerevisiae. Microbiol. Mol. Biol. Rev. 63: 349-404.

Genetic Techniques for Biological Research Corinne A. Michels Copyright © 2002 John Wiley & Sons, Ltd ISBNs: 0-471-89921-6 (Hardback); 0-470-84662-3 (Electronic)

Was this article helpful?

0 0


  • andrea
    How has complementation analysis in medicine?
    7 years ago
  • Esko
    How to calculate no of genes via complementation analysis?
    7 years ago

Post a comment