This page has been archived and is no longer updated
Beyond Mendel: an Evolving view of human genetic disease transmission
Author: Jose Badano
Keywords
Keywords for this Article
Add keywords to your Content
Save
|
Cancel
Share
|
Cancel
Revoke
|
Cancel
Rate & Certify
Rate Me...
Rate Me
!
Comment
Save
|
Cancel
Flag Inappropriate
The Content is
Objectionable
Explicit
Offensive
Inaccurate
Comment
Flag Content
|
Cancel
Delete Content
Reason
Delete
|
Cancel
Close
Full Screen
"� 2002 Nature Publishing Group The rapid identification of genes that are associated with human disease has revolutionized the field of med- ical genetics, providing more accurate diagnostic, prog- nostic and potential therapeutic tools. In addition, an improved understanding of the molecular aetiology of genetic disorders is also altering our perception of dis- ease transmission. The classical model that is used for the discovery of many single-gene disorders is founded on the assumption that the spread of a trait in families is synonymous with the transmission of a single molecu- lar defect. Although some traits are still recognized to be inherited in a monogenic fashion ? with individual alleles segregating into families according to Mendel?s laws ? the number of disorders for which the pheno- type can be satisfactorily explained by mutations at a single locus is now diminishing. It has been suggested that compartmentalizing genetic disorders into monogenic and multifactorial might be an oversimplification 1 . Several studies even indicate that our view of diseases as monogenic might actually be a conceptual artefact 2,3 . Here, we discuss how oligogenic traits were recognized, and what tools can be applied to detect them and dissect their genetic basis. We also review the current molecular models for complex inheritance and illustrate how genetic and mutational data are moving the proposed ?paradigm shift? 4 away from simple models of disease transmission. Despite our improved understanding of multilocus inheritance, the study of true polygenic disorders remains challenging. Nevertheless, the expansion of Mendelian concepts and the construction of theoretical models of increased complexity is an important initial step towards understanding the genetic and molecular basis of multifactorial traits. From the beginning PKU and hyperphenylalaninaemia. PHENYLKETONURIA (PKU) was one of the first genetic disorders for which the biochemical defect was identified long before the advent of polymorphic markers and the ability to carry out genetic analysis in families. A defect in the hepatic enzyme phenylalanine hydroxylase (PAH) was recog- nized during the 1950s (REF. 5), and by 1960, the detec- tion of hyperphenylalaninaemia was offered in neona- tal screening tests. This provided not only an early diagnosis, but also the opportunity to treat the affected individuals 6 (also reviewed in REFS 7,8). Unfortunately ~ 1% of patients did not respond well to the traditional therapy, leading to the realization that, although defects at the PAH locus are present in most PKU cases, both BEYOND MENDEL: AN EVOLVING VIEW OF HUMAN GENETIC DISEASE TRANSMISSION Jose L. Badano* and Nicholas Katsanis* ? Methodological and conceptual advances in human genetics have led to the identification of an impressive number of human disease genes. This wealth of information has also revealed that the traditional distinction between Mendelian and complex disorders might sometimes be blurred. Genetic and mutational data on an increasing number of disorders have illustrated how phenotypic effects can result from the combined action of alleles in many genes. In this review, we discuss how an improved understanding of the genetic basis of multilocus inheritance is catalysing the transition from a segmented view of human genetic disease to a conceptual continuum between Mendelian and complex traits. PHENYLKETONURIA An inborn error of metabolism that is caused by lack of the enzyme PAH that converts phenylalanine to tyrosine. If left untreated, it causes abnormally high phenylalanine levels and severe, progressive mental retardation, but can be prevented by neonatal screening and a low phenylalanine diet from an early age. NATURE REVIEWS | GENETICS VOLUME 3 | OCTOBER 2002 | 779 *Institute of Genetic Medicine, Johns Hopkins University, 2?127 Jefferson Street Building ? Wilmer Eye Institute, 600 North Wolfe Street, Baltimore, Maryland 21287, USA. Correspondence to N.K. e-mail: katsanis@jhmi.edu doi:10.1038/nrg910 REVIEWS HUMAN GENETICS AND DISEASE � 2002 Nature Publishing Group TETRAHYDROBIOPTERIN Phenylalanine hydroxylase (PAH) is an oxygenase that couples an electron from a tetrahydrobiopterin cofactor (BH 4 ) and an oxygen atom to hydroxylate phenylalanine to form tyrosine. Consequently, any defects in BH 4 biosynthesis impair PAH function. 780 | OCTOBER 2002 | VOLUME 3 www.nature.com/reviews/genetics REVIEWS nature of this phenotype, particularly with respect to the substantial phenotypic variability observed in some CF patients. For instance, although CFTR mutations show a degree of correlation with the severity of pancreatic dis- ease, the severity of the pulmonary phenotype ? which is the main cause of mortality ? is difficult to predict (for recent reviews, see REFS 2,14?18). Such realization of the limitations of a pure mono- genic model prompted an evaluation of more complex inheritance schemes. This led to the mapping of a modi- fier locus for the intestinal component of CF in both human and mouse 19,20 . Further phenotypic analysis led to the association of low-expressing mannose-binding lectin (MBL; also known as MBL2) alleles, HLA (human leukocyte antigen) class II polymorphisms, variants in tumour necrosis factor-? (TNFA) and transforming growth factor-?1 (TGFB1) with pulmonary aspects of the disease 21?24 , the correlation of intronic nitric-oxide synthase 1 (NOS1) polymorphisms with variability in the frequency and severity of microbial infections 25 , and the contribution of mucin 1 (Muc1) to the gastrointesti- nal aspects of the CF phenotype in mice 26 (FIG. 1). Recently, further layers of complexity have been dis- covered for both CFTR and its associated phenotype. First, heterozygous CF mutations have been associated with susceptibility to rhinosinusitis, an established mul- tifactorial trait 27 . Second, and perhaps most surprising, a recent study has reported that some patients with a milder CF phenotype do not have any mutations in CFTR. This indicates that the hypothesis that CFTR gene dysfunction is a requisite for the development of CF might not always be true 28 . The emergence of oligogenic disorders The difficulty in establishing a phenotype?genotype correlation, as exemplified by PKU and CF, indicates that, although Mendelian models are useful for identi- fying the primary genetic cause of familial disorders, they might be incomplete models of the true physio- logical and cellular nature of the defect 1,3,4 . Recently, numerous disorders that were characterized initially as being monogenic are proving to be either caused or modulated by the action of a small number of loci (TABLE 1). These disorders are described as ?oligogenic? disorders ? a continuously evolving concept that encompasses a broad spectrum of phenotypes that are neither monogenic nor complex. By contrast to poly- genic traits ? which are thought to result from poorly understood interactions between many genes and the environment ? these oligogenic disorders remain pri- marily genetic in aetiology, but require the synergistic action of mutant alleles at a small number of loci. Many examples indicate that a conceptual contin- uum exists between classical Mendelian and complex traits 29,30 . The position of any given disorder along this continuum depends on three main variables: whether a major locus contributes markedly to the phenotype, the number of loci involved and the extent of environ- mental participation. CF belongs near the beginning of the conceptual continuum, not because of the small number of phenotypic modifiers (there are at least six), allelic heterogeneity and mutations at other loci might also account for the trait in some cases 8 . In 1983, the mapping and cloning of the PAH gene 9 not only con- firmed this biochemical observation, but also showed substantial locus and allelic heterogeneity. In the next decade, the expectation of genetic heterogeneity was substantiated by the discovery of hyperphenylalani- naemia mutations in loci that affect TETRAHYDROBIOPTERIN synthesis or recycling 10 . Consistent with the then emerging view that the genetic complexity in PKU was significantly higher than originally expected for a monogenic disease, several studies reported extensive phenotypic variability, even in the presence of identical genotypes 1,11 . Although the inheritance of mutant alle- les follows a Mendelian segregation pattern, the identi- fication of mutations in any of the genes that cause hyperphenylalaninaemia cannot predict the phenotype of the patient, indicating that genetic factors and the environment might be important modulating agents. From biochemical models to pure genetics Cystic fibrosis. By contrast to PKU ? in which the defect was first defined (at least in part) by biochemical means ? cystic fibrosis (CF) represents an early example of the use of pure genetic models to identify the mutated gene. On the basis of the observed autosomal-recessive inheri- tance in families, the gene CFTR (cystic fibrosis trans- membrane conductance regulator) was first mapped in humans to chromosome 7q31.2 (REF. 12). The CFTR gene was cloned 13 , fuelling speculation that mutation analyses might be sufficient to predict the clinical outcome of patients. The analyses of CFTR mutations in large and ethnically diverse cohorts indicated that the initial hypothesis was an oversimplification of the true genetic CFTR Gastrointestinal phenotype Microbial infections Rhinosinusitis susceptibility Severity of pulmonary phenotype Meconium ileus HLAII TNFA TGFB1 MBL2 CFM1 Muc1 NOS1 Figure 1 | Complexity in monogenic diseases. Mutations in CFTR almost always cause the CF phenotype. Owing to modification effects by other genetic factors, the presence and nature of mutations at the CFTR locus cannot predict what the phenotypic manifestation of the disease will be. Therefore, although CF is considered a Mendelian recessive disease, the phenotype in each patient depends on a discrete number of alleles at different loci. Meconium ileus describes the obstruction at birth of the small and/or large intestine (ileus) with the first faecal excretion (meconium). CF, cystic fibrosis; CFTR, cystic fibrosis transmembrane conductance regulator; CFM1, cystic fibrosis modifier 1; HLA-II, MHC class II antigen; MBL2, mannose-binding lectin (protein C) 2; NOS1, nitric oxide synthase 1; TGFB1, transforming growth factor-?1; TNFA, tumour necrosis factor-? encoding gene. � 2002 Nature Publishing Group NATURE REVIEWS | GENETICS VOLUME 3 | OCTOBER 2002 | 781 REVIEWS number of genes, with eight mapped potential loci and a handful of proposed candidate genes (for a recent review, see REF. 31). The far end of the spectrum might contain traits that influence behaviour or cognition, the genetic basis of which is too poorly understood (and often controversial) to quantify. but owing to the prevalence of a major locus, CFTR, which accounts for most of this phenotype. Schizophrenia, on the other hand, resides in the middle of the continuum because the substantial genetic pre- disposition (40?50% phenotypic concordance in twin studies) is probably conferred by a relatively small Table 1 | Examples of human oligogenic disorders Syndrome or trait OMIM number Primary locus Secondary locus Effect References Non-syndromic recessive 220290 GJB2 GJB6 Digenic 102 deafness (DFNB1) Non-syndromic recessive 605429 DFNB26 DFNM1 Suppressor 103 deafness (DFNB26) Usher syndrome type I 276903 USH3 MyoVIIA Synergistic 104 Non-syndromic dominant deafness 601842 DFNA12 DFNA2 Digenic/additive 105 Waardenburg syndrome 103470 MITF TYR Digenic 106 type II and ocular albinism Retinitis pigmentosa 180721 RDS ROM1 Digenic 44 Limb-girdle muscular dystrophy 2A 114240 CAPN3 Postulated Penetrance/expressivity 107 Dysfibrinogenaemia causing 134820 FGA FGG Digenic 108 recurrent thrombosis Bardet?Biedl syndrome 209900 BBS6/BBS2 BBS2/BBS6 Digenic ?triallelic? 50 BBS2/BBS4 BBS4/BBS2 Digenic ?triallelic? 46 Hirschsprung disease 142623 RET GDNF Digenic 85 RET 3p21;19q12 Penetrance/risk 59 Junctional epidermolysis bullosa 150310 LAMB3 COL17A1 Digenic 61 Cystinuria type III 600918 SLC7A9 SLC3A1 Digenic? 109 Becker muscular dystrophy 159991 DMD MYF6 Severity modifier 110 Breast cancer 175100 BRCA APC Modifier/risk 111 Spinal muscular atrophy 603011 SMN1 H4F5 Candidate severity modifier 112 Cystic fibrosis 603855 CFTR CFM1 Suppressor 19 Breast and ovarian cancer 113705 BRCA1 HRAS1 Penetrance/risk 113 Familial amyotrophic lateral 147450 SOD1 CNTF Severity modifier 33 sclerosis Familial hypercholesterolaemia 143890 LDLR 13q Suppressor 114 Familial Mediterranean fever 249100 MEFV SAA1 Pleiotropy 115 Maternally inherited deafness 561000 12S ribosomal D8S277 Penetrance 116 Melanoma 600160 CDKN2A MC1R Penetrance/risk 117 Van der Woude syndrome 604547 VWS 17p11.2 Penetrance 118 Type I von Willebrand disease 601628 VWF ABO blood group Penetrance 119 VWF Galgt2 Modifier 120 Nephrotic syndrome 256300 NPHS1 NPHS2 Digenic 66 Autosomal-dominant glaucoma 137750 MYOC CYP1B1 Severity modifier 74 Congenital disorder of 212065 PMM2 ALG6 Severity modifier 94 glycosylation type Ia Alzheimer?s disease 104300 APP TGFB1 Severity modifier 121 Familial adenomatous polyposis 175100 Apc Mom1, Cox2, cPLA2 Protection 38,122,123 Alagille syndrome 118450 Jag1 Notch2 Digenic (mouse model) 90 ALG6, dolichyl-P-Glc:Man 9 GlcNAc 2 -PP-dolichylglucosyltransferase; APC/Apc, adenomatosis polyposis coli; APP, amyloid-?(A4) precursor protein; BBS2/4/6, Bardet?Biedl syndrome 2/4/6; BRCA1, breast cancer 1; CAPN3, calpain 3, (p94); CDKN2A, cyclin-dependent kinase inhibitor 2A; CFM1, cystic fibrosis modifier 1; CNTF, ciliary neurotrophic factor; CFTR, cystic-fibrosis transmembrane-conductance regulator; COL17A1, collagen, type XVII, ?1; Cox2, cyclooxygenase 2; cPLA2, phospholipase A2; CYP1B1, cytochrome P450, subfamily I (dioxin-inducible), polypeptide 1; D8S277, Diamond?Blackfan anaemia 2; DFNA2, potassium voltage-gated channel, KQT-like subfamily, member 4; DFNA12, tectorin-?; DFNB26, deafness, autosomal recessive 26; DFNM1, deafness modifier 1; DMD, dystrophin; FGA, fibrinogen, A ?-polypeptide; FGG, fibrinogen, ?-polypeptide; Galgt2, UDP-N-acetyl-?-D-galactosamine:(N-acetylneuraminyl)-galactosyl-N-acetylglucosaminylpolypeptide-?1, 4-N-acetylgalactosaminyltransferase; GDNF, glial-cell-derived neurotrophic factor; GJB, gap-junction protein ?; H4F5, small EDRK-rich factor 1A (telomeric); HRAS1, v-Ha-ras Harvey rat sarcoma viral oncogene homologue; Jag1, jagged 1; LAMB3, laminin- ?3; LDLR, low-density-lipoprotein receptor; MC1R, melanocortin 1 receptor; MEFV, Mediterranean fever; MITF, microphthalmia-associated transcription factor; Mom1, phospholipase A2, group IIA; MYF6, myogenic factor 6; MyoVIIA, deafness modifier 1; MYOC, myocilin, trabecular meshwork inducible glucocorticoid response; Notch2, Notch gene homologue 2; NPHS, nephrosis 1/2; PMM2, phosphomannomutase 2; RDS, retinal degeneration, slow; RET, ret proto-oncogene; ROM1, retinal-outer-segment membrane protein 1; SAA1, serum amyloid A1; SLC7A9/3A1, solute carn�r family 7, member 9/solute carrier family 3, member 1; SMN1, survival of motor neuron 1; SOD1, superoxide dismutase; TGFB1, transforming growth factor, ?1; TYR, tyrosinase; USH3, Usher syndrome 3; VWF, von Willebrand factor; VWS, Van der Woude syndrome. � 2002 Nature Publishing Group 782 | OCTOBER 2002 | VOLUME 3 www.nature.com/reviews/genetics REVIEWS Phenotypic differences in animal models. Recapitulating human phenotypes in experimental animal models allows the human mutation to be isolated and examined in a fixed genetic background. This has been a powerful tool for both establishing oligogenicity and mapping the loci that are involved in this phenomenon. FAMILIAL ADENOMATOUS POLYPOSIS (FAP) is caused by mutations in the adenomatous polyposis coli (APC) gene and is a dis- order that is thought to lie early in the continuum of phenotypic causality, representing a classic example of such modelling. The phenotype of the dominant ENU (ethyl nitrosourea mouse mutant Min (multiple intesti- nal neoplasia), caused by a point mutation in the Apc gene 35 , was shown to be modulated by a second locus, Mom1 (modifier of Min), with respect to the number of tumours that developed 36,37 . Genetic analysis of the modification effect on the Min locus led to the mapping of Mom1 to mouse chromosome 4 (REF. 38;see REF. 17 for a more extensive description). Another example of this approach is the recent cloning of a modifier locus for the phenotype caused by the mouse tubby (tub) mutation. This spontaneous, autoso- mal-recessive mutation is characterized by obesity, insulin resistance and sensory defects, including retinal degenera- tion and hearing loss. The tub phenotype was reported in 1990 (REF. 39) and positional cloning identified it as being caused by mutations in the Tub gene, which encodes a novel protein of unknown function 40,41 .However,geneti- cally controlled modifications of the tub phenotype were observed, in which the modifier conferred protection against hearing loss. QUANTITATIVE TRAIT LOCUS (QTL) analysis then mapped the modifier of tubby hearing 1 (moth1) to mouse chromosome 2 and showed that wild-type Moth1 alleles, present in the AKR/J, CAST/EiJ and 129P2/OlaHsd strains, provided this protection to tub mice in a domi- nant fashion 42 . On the basis of this information, the Moth1 gene was cloned and shown to correspond to the microtubule-associated protein 1a (Map1a ) gene 43 . Disparities between mutations and Mendelian models. By contrast to directed modelling studies that revealed oligogenicity, many of the oligogenic traits that were initially thought to be purely monogenic ? such as cases in which mutations were found to contradict the classical Mendelian model ? have been detected by fortuitous means. RETINITIS PIGMENTOSA (RP), a genetically and clinically heterogeneous disease that can be inher- ited as an autosomal-dominant, autosomal-recessive or X-linked trait, represents the first clear example of how an expansion of the Mendelian model has resolved conflicting mutational and genetic data. Using a digenic model of disease transmission, Kajiwara and colleagues showed that heterozygous mutations in both the retinal outer segment membrane protein 1 (ROM1) gene and the peripherin/retinal degeneration slow (RDS) gene were required in some seemingly dominant RP pedi- grees to cause disease 44 (FIG. 2b). In three families, the RP phenotype was shown to segregate with a L185P mis- sense mutation in RDS and a 1-bp deletion at codon 114 in ROM1, with individuals carrying either of the two mutations being asymptomatic 45 . Establishing oligogenicity It is not surprising that our understanding of genetic disorders decreases precipitously the further we deviate from Mendelian models. However, the application of statistical and molecular tools that were developed using monogenic systems has shown some success in the modelling and/or cloning of genes that are involved in oligogenic traits. In the oligogenic disorders that have been character- ized, several complementary approaches have been applied to show oligogenicity and to clone the loci responsible for the disease. Methodologically, we broadly recognize four lines of investigation: pheno- type?genotype correlations, phenotypic differences in an animal model of the disease that are dependent on the genetic background, the identification of bona fide mutations that do not conform to monogenic inheri- tance and the establishment of linkage to more than one locus or the failure to detect linkage using Mendelian models. Phenotype?genotype correlations. After the discovery of a disease gene, a natural first step is to investigate the mutational spectrum in large patient cohorts. This leads to attempts to correlate specific mutations (or classes of mutation, such as nonsense versus mis- sense) with various phenotypic aspects, such as sever- ity and age of onset. The frequent inability to draw conclusions from such studies prompts the expan- sion of the monogenic model of disease transmission to account for other factors. Difficulties in predicting the phenotype of PKU and CF patients from muta- tional data are early examples of observations that now abound in the literature. For instance, in familial amyotrophic lateral sclerosis (FALS) ? a neurologi- cal disorder that is transmitted primarily as a domi- nant trait 32 ? a family was reported in which the mother, son and daughter carried the same V148G mutation in the gene encoding copper/zinc superox- ide dismutase 1 (SOD1). The mother developed FALS at the age of 54 and died at 55, the daughter was asymptomatic at 35, but the son developed severe FALS at the age of 25 and died within 11 months. This prompted Giess and colleagues to screen for candidate modifiers that might have an impact on the age of onset. They reported that the affected son also carried a homozygous null mutation in the ciliary neurotrophic factor (CNTF) gene, thereby implicat- ing CNTF as a modifier of FALS 33 (FIG. 2a). Naturally, because these observations were based on the find- ings from a single family, it remains unclear how commonly CNTF modulates the FALS phenotype. Interestingly, the same CNTF mutation had been reported previously in Japanese patients with various neurological disorders. However, as no correlation with the disease could be drawn on the basis of the Mendelian model of transmission, the mutation was concluded not to be causally related to the disease 34 . In retrospect, we propose that the neurological dis- ease in these Japanese patients might be attributable to mutations in CNTF and other loci. FAMILIAL ADENOMATOUS POLYPOSIS (FAP). The development of numerous adomatous polyps in the colon that might progress to carcinomas. QUANTITATIVE TRAIT LOCUS (QTL). A genetic locus or chromosomal region that contributes to variability in complex quantitative traits (such as height or body weight), as identified by statistical analysis. Quantitative traits are typically affected by several genes and by the environment. RETINITIS PIGMENTOSA (RP). A group of both clinically and genetically heterogeneous hereditary retinal degeneration disorders that are caused by the death of both rod and cone photoreceptors, leading to a complete loss of vision. � 2002 Nature Publishing Group NATURE REVIEWS | GENETICS VOLUME 3 | OCTOBER 2002 | 783 REVIEWS both of his children, only one of whom was affected, although haplotype analysis indicated that both siblings had inherited the same maternal chromosome (FIG. 3a). Therefore, even if the mother carried an undetected mutation, she would have to have transmitted it to both siblings, which would lead to an unaffected individual carrying two heterozygous BBS6 mutations. This would indicate that, in some families, more than two mutant alleles might be required to manifest BBS. Indeed, the affected sibling was homozygous across the critical inter- val of another BBS gene, BBS2, indicating that this pedi- gree might segregate two BBS2 mutations and one BBS6 mutation 48 . The positional cloning of BBS2 (REF. 49) con- firmed this hypothesis. Consistent with the haplotype data, a homozygous Y24X BBS2 mutation was detected in the affected but not the unaffected sibling 50 , indicating the existence of another variation of digenic inheritance, ?triallelic? inheritance, which initially seemed to be trans- mitted as a recessive trait 50 . Consistent with this oli- gogenic mode of disease transmission, three mutations at two loci were found in three more BBS pedigrees, including one family in which the patient inherited two nonsense BBS2 mutations and a nonsense BBS6 muta- tion, and his unaffected brother had two BBS2 nonsense mutations but was wild type for BBS6 (REF. 50; FIG. 3b). Linkage studies. Statistical tools for modelling the genetic interactions between loci and environmental factors are proliferating rapidly. Although mathematical analyses of oligogenicity are beyond the scope of this discussion (see REFS 51,52 for in-depth analyses), it is important to recognize that the modified use of tradi- tional linkage approaches remains a useful tool for the study of oligogenic diseases, especially if a major locus that contributes greatly to the phenotype is known. This is exemplified by recent developments in Hirschsprung disease (HSCR), a congenital disorder that is characterized by the variable absence of ENTERIC GANGLIA. On the basis of the extent of aganglionosis, two main phenotypic groups can be distinguished: short-segment HSCR (S-HSCR) and the more severe long-segment HSCR (L-HSCR) 53 . Autosomal-dominant inheritance with incomplete PENETRANCE has been proposed for L-HSCR, whereas complex inheritance that involves an autosomal-recessive trait has been observed in S-HSCR. Oligogenicity has been established in both HSCR vari- ants by virtue of several factors: a recurrence risk that varies from 3 to 25%, depending on the length of agan- glionosis and the sex of the patient; HERITABILITY values close to 100%, which indicates an exclusively genetic basis; significant clinical variability and reduced pene- trance; and non-random association of hypomorphic changes in the endothelin receptor type B (EDNRB) with rearranged during transfection (RET) polymor- phisms and HSCR 54 (for a comprehensive discussion; see REF. 55). So far, a combination of linkage, positional- cloning studies and functional candidate gene analyses has identified eight HSCR genes (for a review, see REF. 56), of which the proto-oncogene RET 57,58 is thought to be the main predisposing locus, particularly in families with a high incidence of L-HSCR 59 . Depending on the level of influence that each of the two loci exerts on the phenotype, digenic inheritance can be considered as either synergistic or modifying. Typically, those traits that segregate in families in a dom- inant fashion probably manifest a synergistic effect, whereas recessive disorders might present a modifying effect. Naturally, variation between these two models are likely to manifest in an allele-and-context-dependent manner. A recent example of a recessive trait, the non- Mendelian inheritance of which was established by mutational data, is BARDET?BIEDL SYNDROME (BBS), a geneti- cally heterogeneous disease that is thought to be caused by recessive mutations in one of at least six genes. Mutational and HAPLOTYPE ANALYSIS of BBS6, the first BBS gene to be cloned 46,47 , indicated that some mutations did not conform to the expected recessive transmission. In one CONSANGUINEOUS pedigree in particular, a heterozy- gous A242S mutation was transmitted from the father to BARDET?BIEDL SYNDROME (BBS). A rare and genetically heterogeneous disorder that is characterized primarily by obesity, retinal dystrophy, polydactyly, hypogenitalism, learning difficulties and renal malformations. HAPLOTYPE ANALYSIS The study of the pattern of descent of a combination of alleles at different sites on a single chromosome (known as a haplotype). It is used for the identification of recombination events between markers and traits during linkage studies, thereby establishing the boundaries of the location of a phenotype-associated locus. Age of onset 56 Age of onset 43 CNTF +/? SOD +/+ CNTF +/? SOD +/+ CNTF +/? SOD +/+ CNTF +/+ SOD +/+ CNTF +/+ SOD +/? Age of onset 54 CNTF ?/? SOD +/? Age of onset 25 CNTF +/? SOD +/? Age of onset 55 a ROM +/+ RDS +/+ Non-affected ROM +/? RDS +/+ Non-affected ROM +/? RDS +/? Affected ROM +/+ RDS +/+ Non affected ROM +/+ RDS +/? Non-affected ROM +/? RDS +/? Affected b Figure 2 | Idealized pedigrees showing examples of complex phenotypic modulation. a | A CNTF-null allele modulates the age of onset of the dominantly transmitted disease amyotrophic lateral sclerosis (ALS). Mutations in both SOD1 and CNTF lead to early-onset ALS (age 25) and death within 11 months (the third-generation male in the diagram). The third- generation female with a SOD1 mutation but no CNTF mutation did not present with the disorder until the age of 54. b | A pedigree of digenic inheritance showing how retinitis pigmentosa occurs only in individuals who have inherited one mutation in each of ROM1 and RDS. Heterozygotes for either mutant allele are asymptomatic. CNTF, ciliary neurotrophic factor; RDS, retinal degeneration slow; ROM1, retinal outer segment membrane protein 1; SOD1, superoxide dismutase 1. Circles indicate females; squares indicate males; diagonal red line across symbol indicates lethality. � 2002 Nature Publishing Group 784 | OCTOBER 2002 | VOLUME 3 www.nature.com/reviews/genetics REVIEWS expanded models of disease inheritance that account for reduced penetrance and phenotypic variability and the ability of these models to genetically map loci involved in oligogenic diseases ? a first step towards identifying genes that underly them. More importantly, the estab- lishment of non-Mendelian models has caused a change of perception in human genetics which, in turn, has accelerated the discovery of oligogenic traits. Although genetic studies have clearly been useful, most of the success in this field has relied on the avail- ability of many genes that result in the same or similar phenotypes in an apparently monogenic way. Therefore, it is not surprising that the genetic basis of non- Mendelian inheritance has been demonstrated more frequently in genetically heterogeneous disorders. The key to the discovery of digenic inheritance in RP and BBS was not only the identification of mutations that conflicted with the familial transmission of the trait, but also the availability of several genes, each of which had independently been shown to cause the same pheno- type. Given that human patient cohorts are not always sufficiently informative to conduct genetic studies, these ?candidate-gene? approaches (once complex inheritance has been established) show great promise. This is exem- plified by junctional epidermolysis bullosa (JEB), a group of recessive disorders of the skin that range in severity from lethal (JEB Herlitz) to milder forms, such as generalized atrophic benign epidermolysis bullosa (GABEB). In one family, the proband presented with a severe phenotype that had aspects of both JEB Herlitz and GABEB, which led the investigators to query the genes known to cause these disorders, including colla- gen XVII (COL17A1) and the ?3-subunit of laminin 3 (LAMB3). This resulted in the detection of two COL17A1 nonsense mutations (L855X and R1226X) and a heterozygous LAMB3 mutation (R635X) in the patient 61 . Recently, digenic inheritance has been proposed for NEPHROTIC SYNDROME. Nephrin (NPHS1) and podocin (NPHS2) have been shown to cause two types of nephrotic syndrome, the Finnish congenital nephrotic syndrome (CNF) and autosomal-recessive familial focal segmental glomerulosclerosis (SRN1) 62?64 . Functional studies have indicated that these two proteins interact physically 65 , and the recognition that nephrotic syn- drome has substantial clinical and genetic variability prompted Koziell and colleagues to screen both NPHS1 and NPHS2 in patients with this disorder 66 . Consistent with the expectation of oligogenicity, four patients were identified who carried mutations in both genes in a model similar to triallelic inheritance 66 . The nature and effect of oligogenic mutations Although mapping and cloning loci is a major chal- lenge, defining oligogenic mutations is often substan- tially more problematic. A key hurdle that can prove difficult to overcome is establishing the specific alter- ations in the DNA sequence as deleterious. Fortunately, the effect of ?major? mutations, such as deletions, insertions, splice-site changes and nonsense alterations, is relatively easy to both predict and test. The non-Mendelian transmission of HSCR has hin- dered the identification of predisposing modifier loci by conventional linkage approaches. When both PARAMETRIC and NON-PARAMETRIC LINKAGE (NPL) studies were carried out on a group of 12 L-HSCR families, weak LOD SCORES and NPL values were observed on 9q31. However, on the basis of the hypothesis that only milder RET mutations could be associated with another locus, families were categorized according to the RET mutational data. Significant linkage on 9q31 was detected when families with potentially weak RET mutations were analysed independently 60 , indicating that mild RET alleles, in conjunction with alleles at an unknown gene on chromosome 9, might be required for pathogenesis. The mode of inheritance in S-HSCR has proved to be more complex than L-HSCR, and has required fur- ther adjustments to the linkage strategies. Recently, the application of model-free linkage, without assumptions about the numbers and inheritance mode of segregating factors, showed that a three-locus segregation was both necessary and sufficient to manifest S-HSCR, with RET being the main locus, and that the transmission of sus- ceptibility alleles was additive 59 . The impact of oligogenicity on genetic research The inheritance patterns observed in disorders such as Hirschsprung disease illustrate the power of both CONSANGUINEOUS Descended from a recent common ancestor. ENTERIC GANGLIA Parasympathetic mass of nerve tissue (ganglia) in the colon. PENETRANCE The proportion of affected individuals among the carriers of a particular genotype. If all individuals who have a disease genotype show the disease phenotype, then the disease is said to be ??completely penetrant??. HERITABILITY The proportion of the variation in a given characteristic or state that can be attributed to genetic factors. PARAMETRIC LINKAGE Parametric analyses are statistical tests for linkage that use assumptions such as mode of transmission, allele frequencies and penetrance. 01 Y24X wt 02 a NFB14-BBS2 Y24X wt 03 Y24X Y24X 04 wt wt 01 wt wt 02 NFB14-BBS6 wt A242S 03 wt A242S 04 wt A242S 01 wt Q59X wt wt 02 b AR259-BBS2 wt wt Y24X wt 03 wt Q59X Y24X wt 04 wt wt wt wt 05 wt Q59X Y24X wt 01 Q147X wt 02 AR259-BBS6 wt wt 03 Q147X wt 04 wt wt 05 wt wt Figure 3 | Triallelic inheritance. a | In the consanguineous pedigree NFB14 both the affected (03) and the unaffected (04) individuals carry the same mutation (A242S) in the Bardet?Biedl syndrome gene, BBS6. Only the affected sibling is homozygous for a nonsense mutation (Y24X; X indicates a stop codon) in BBS2. The double line that links the couple in the first generation indicate consanguineity. b | Three mutations at two loci are necessary for pathogenesis in this pedigree, as the affected sibling (03) has three nonsense mutations (Q147X in BBS6, and Y24X and Q59X in BBS2) and the unaffected sibling (05) has two nonsense BBS2 mutations, but is wild-type for BBS6. � 2002 Nature Publishing Group NATURE REVIEWS | GENETICS VOLUME 3 | OCTOBER 2002 | 785 REVIEWS probably a modifier locus. However, in the absence of functional data for some variants and none for others it is difficult to discriminate whether some of these alleles are pathogenic are in LINKAGE DISEQUILIBRIUM with the true mutation or are the product of random GENETIC DRIFT in the population. Furthermore, evaluation of the effect of particular alleles in isolation might provide limited information, as the true biochemical defect could be unmasked only by modelling the synergistic action of several mutations acting both in cis and in trans. Examining oligogenic alleles in model organisms, such as the mouse, can overcome some of these difficul- ties, particularly because the effect of an allele can be tested in a genetically homogeneous background. Such experimental animal models are particularly useful when some information on the gene function is avail- able or when a tissue/organ phenotype has been estab- lished. For instance, the protection afforded by moth1 in the Tub hearing phenotype seems to be attributable to a shorter Ala-Pro repeat region in Map1a, whereby the protective strains 129P2/OlaHsd and CAST/Ei contain a Map1a protein with two and five repeat subunits fewer than the B6 strain, respectively 43 . Genetic and sequenc- ing data alone cannot distinguish between a mutation at this site versus linkage disequilibrium between a poly- morphism and the actual modifier mutation, which might or might not map to the same gene. However, partial rescue of the Tub hearing phenotype with a Map1a transgene containing the 129P2/OlaHsd allele, provides strong evidence that Map1a is moth1.Co- immunoprecipitation studies also indicated that the observed differences between the various Map1a alleles might result in different binding affinities for specific proteins 43 , although further work is required to substan- tiate this intriguing hypothesis. Molecular mechanisms of oligogenicity By contrast to genetic modelling of oligogenicity, of which there are now numerous examples (TABLE 1), the paucity of functional data for many of the genes that cause human disease means that the molecular basis of such phenomena is poorly understood. One of the best-studied illustrations for the molecu- lar basis of complex inheritance is the digenic interac- tion of ROM1 and RDS that causes RP. These two proteins form homodimers, which in turn interact to form tetrameric complexes 79 that are important for the structural integrity of photoreceptors in the retina 80,81 . The digenic RDS mutation prevents the formation of functional RDS?RDS homocomplexes 82,83 , with the null ROM1 mutation further decreasing the amount of functional complexes available, and is postulated to lead to photoreceptor degeneration 82,83 . Other examples in which synergistic mutations at discrete loci underlie a direct interaction between their encoded proteins illustrates the same principle. The genetic interaction of NPHS1 and NPHS2 in nephrotic syndrome could be attributed to the direct interaction of the carboxyl terminus of NPHS2 with NPHS1 (REFS 65,84). However, the effect of the muta- tions on the stability and function of this complex By contrast, predisposing or protecting alleles, the effect of which on the phenotype is often subtle, is considerably more difficult. Some mutations might be mild coding sequence alterations, might lie in the reg- ulatory regions of transcripts such as 5? and 3? termi- nal regions and introns or might lie in control elements that are several kilobases away from the transcript. In the absence of alternative means of assaying the mutagenic effect of an allele, it can be challenging to conclude using genetic and population data that a particular gene is causal to the phenotypic modulation. The digenic inheritance in BBS exemplifies this problem. Shortly after the discovery of triallelic inheri- tance between BBS2 and BBS6, the third gene for this disorder, BBS4, was identified 67 . Subsequent mutational and genetic analyses indicate that mutations in BBS4 also interact with mutations at other loci, including the possibility that two BBS4 mutant alleles and two BBS2 mutant alleles are necessary for pathogenesis in some pedigrees. One patient had two BBS2 T560I mutations and two BBS4 A364E mutations, whereas his asympto- matic mother and brother were T560I/T560I, A364E/wt. In another three pedigrees, only a single BBS4 mutant allele was found (L327P, N165H and S457I), and each pedigree was excluded by haplotype analysis from having a second mutation in BBS4 68 .In the absence of functional information, determining which, if any, of these alterations are pathogenic is prob- lematic. It is not possible to distinguish between rare benign variants and subtle mutations, or, in the absence of complete knowledge of all key regulatory elements in BBS4, can it be claimed that all potential pathogenic variations have been detected. GLAUCOMA is another example that shows the diffi- culty in linking DNA variations with pathogenic causal- ity. Mutations in several genes have been associated with different types of glaucoma, including myocilin muta- tions (MYOC) in juvenile open-angle glaucoma 69 (JOAG) and CYP1B1 (cytochrome P450, subfamily I, polypeptide I) in primary congenital glaucoma 70 (PCG). Oligogenicity in glaucoma has been suspected because of substantial inter- and intrafamilial pheno- typic variability 71 , and because of the identification of non-expressing individuals who carry two mutant alle- les without developing the disease 72,73 . Recently, a pedi- gree was reported in which glaucoma segregated in an autosomal-dominant fashion. In this family, patients with a heterozygous G399V mutation in MYOC and a heterozygous R368H mutation in CYP1B1 manifested the disease at a mean age of 27 years, whereas individu- als with only the heterozygous MYOC mutation devel- oped glaucoma at a mean age of 51 years (REF.74). Moreover, a V432L variant in CYP1B1, which is thought to be a polymorphism, was also found in the same fam- ily. The leucine allele has been proposed to affect the ability of the protein to hydroxylate 17?-oestradiol, a target of CYP1B1 some studies have reported a decrease in catalytic efficiency 75?77 , whereas others have reported the converse 78 . Collectively, the MYOC and CYP1B1 genetic and mutational data indicate that CYP1B1 is NON-PARAMETRIC LINKAGE Non-parametric approaches are statistical procedures that are not based on models, or assumptions pertaining to the distribution of the quantitative trait. LOD SCORE (Base 10 ?logarithm of the odds? or ?log-odds?). A method of hypothesis testing. The logarithm of the ratio between likelihoods under the null and alternative hypotheses. NEPHROTIC SYNDROME Malfunction of the renal glomerular filtration barrier (a structure in the glomerulus that is responsible for protein filtration) that lead to the loss of plasma proteins. GLAUCOMA The abnormally elevated pressure in the liquid that fills the anterior part of the eye (the aqueous humour). LINKAGE DISEQUILIBRIUM The condition in which the frequency of a particular haplotype for two loci is significantly greater than that expected from the product of the observed allelic frequencies at each locus. GENETIC DRIFT Random fluctuations in the allele and, less commonly, the phenotype frequencies, as genes are transmitted from one generation to the next. � 2002 Nature Publishing Group 786 | OCTOBER 2002 | VOLUME 3 www.nature.com/reviews/genetics REVIEWS These examples illustrate the general principle of non- allelic non-complementation, whereby mutations in two different genes can behave as alleles of the same locus by causing or exacerbating the same phenotype. This phe- nomenon has been established in various organisms and in numerous physiological processes, including cuticle development, transcriptional regulation, neuron out- growth and cytoskeletal motility 91 .The two main models that explain non-allelic non-complementation are the dosage model and the poison model 92,93 . According to the dosage model, the simultaneous decrease in dosage at two loci is required to manifest the phenotype (FIG. 4a,b). In the requires clarification. Finally, receptor?ligand rela- tionships might also underlie the physiological basis for some types of complex inheritance. Oligogenicity in some Hirschsprung families is attributed to muta- tions in both RET and the glial-cell-derived neu- rotrophic factor (GDNF)gene 85 ; with RET forming part of a multisubunit receptor of GDNF 86,87 . Likewise, phenotypic modulation in ALAGILLE SYNDROME, a disease caused by HAPLOINSUFFICIENCY of the jagged (JAG1) gene product 88,89 , has been attributed in part to mutations in NOTCH2, which encodes a receptor for JAG1 (REF. 90). ALAGILLE SYNDROME A dominantly inherited disorder that is characterized primarily by a scarcity of bile ducts in the liver. Other features include heart, eye, kidney and skeletal abnormalities, as well as defects in the central nervous system. HAPLOINSUFFICIENCY A gene dosage effect that occurs when a diploid requires both functional copies of a gene for a wild-type phenotype. An organism that is heterozygous for a haploinsufficient locus does not have a wild-type phenotype. ab AB AB AB A * * * B A * Bor D A B B B B C A B C D A B C D A B C D A B C D A B C D A B C D A B C D A B C A Function C A C D D D D Normal A B C D A B C D A B C A B C A B C D D D D A B C Normal Affected A +/? D +/? D A B C D A B C D A B C Normal A B C D D B C A D B C A D B C A D B C A D B A C D B A C D B A C c A +/+ D +/+ A +/+ D +/? A +/? D +/+ Figure 4 | Models of non-allelic complementation. a, b | The direct-interaction dosage model. a | Mutations at one locus (mutated proteins are indicated by asterisks) are not sufficient to disrupt the formation of the complex between proteins A and B, although the strength of the interaction might be reduced (dashed line). A further mutation in protein B causes disruption of the complex (red cross), resulting in a detectable phenotype. b | A similar model involving proteins B and D, which are members of the same multi-subunit complex but do not interact directly. c | The poison model. Mutations in protein A disrupt the complex, although enough functional units remain to maintain function. A further mutation (or mutations) in protein D disrupts more units, resulting in the disruption of a physiological process and the generation of a cellular phenotype. � 2002 Nature Publishing Group NATURE REVIEWS | GENETICS VOLUME 3 | OCTOBER 2002 | 787 REVIEWS Consequently, the diagnosis and management of genetic diseases will continue to be a case-by-case process and to rely heavily on epidemiological data for each disease and its associated alleles. Recognition of these limitations is the first step towards a solution. A better understanding of the mole- cular mechanisms of oligogenicity will probably enhance our ability to make accurate genotype-based phenotypic predictions and to then estimate more clearly the effect of the environment. Furthermore, the identification of loci, the concurrent dysfunction of which leads to a specific phenotype, represents an important tool for research in protein function and cel- lular pathways. poison model, a mutation impairs the protein com- plex, while retaining function, probably because enough functional complexes remain available to the cell. A second mutation at another protein of the same complex further disrupts the already reduced number of normal complexes and leads to an observable effect (FIG. 4c). Ligand?receptor relationships (such as JAG1?NOTCH2) might conform to the dosage model, whereas the poison model could describe phenomena such as digenic RP. Molecular models of oligogenicity need not be con- fined to proteins that interact directly. In Caenorhabditis elegans, non-allelic complementation has been observed between UNC-13 and synaptobrevin, two proteins that bind to the same complex but that are not directly asso- ciated with each other 91 . In other models, two proteins might act at different stages of the same pathway; in such cases, a mutation in each protein contributes quantita- tively to the progressive dysfunction of the pathway until a crucial threshold is reached and a disease phenotype is observed (FIG. 5). This is exemplified by some congenital disorders of glycosylation, whereby mutations in ALG6 and PMM2, each encoding an enzyme involved in a dif- ferent part of the post-translational-modification process, might be required to manifest the more severe form of the disease 94 . Finally, mutations at different posi- tions of the same protein network, or in independent but synergistic networks, might also result in oligogenic phe- notypes. There is no doubt that, as our understanding of the nature and composition of protein networks and sig- nalling pathways increases (for example, see REFS 95,96), new examples will surface, and our molecular models to describe oligogenicity will improve in both accuracy and sophistication. The road ahead Oligogenic disorders represent both a unique challenge and an opportunity. The realization that many of the genetic disorders that were described previously as monogenic are in fact the product of defects at a small number of loci creates conceptual and practical prob- lems. First, it requires the reorganization of our approach to studying human genetic disorders, starting by dissoci- ating the mode of inheritance in families (that is, Mendelian segregation of traits) from the mode of inher- itance of specific disease alleles. As most mutations prob- ably exert a quantitative effect on the phenotype, labelling mutant alleles as dominant or recessive is often an over- simplification. Similarly, diagnoses and patient management can be impeded by oligogenic phenomena 97 . In most cases, although the likelihood of disease onset that is attrib- utable to a mutant allele can be predicted, the pheno- typic outcome cannot. For example, the onset of FALS at age 25 or age 54 are two different scenarios, each having a profound impact on the life choices of the patient (FIG. 2a). In other cases, the absence of CF muta- tions might not be sufficient to exclude CF or, on the contrary, the presence of two mutations in one gene might not always predict the onset of disease, as is the case for some triallelic BBS pedigrees (FIG. 3b). Wild type or B +/? a A C Signal B ?/? A C Signal B ?/? ; S +/? A C Reduced signal B ?/? ; S ?/? A C No signal b A C A C A C A C H D Signal No signal D Signal No signal BB ?/? B H B H Signal Signal Reduced signal Reduced signal BSSBBBS Figure 5 | Idealized pathway-complementation schemes. a | Signal transduction between protein A and protein C (first column) is disrupted by the homozygous loss of function of protein B (second column). The substitution of the function of protein B by S rescues the pathway. Heterozygous mutations in S decrease the efficiency of the pathway and might cause an observable phenotype (third column). By contrast, loss of S causes a complete loss of signal (last column). The histogram illustrates the relative strength of the signal that is associated with each genotype. b | Dosage-dependent effects on members of the same pathway. Null alleles of any of proteins A?D cause a loss of signal (a mutation in B is shown in the second column). However, hypomorphic (H) mutations in B might maintain the signal, albeit at a reduced strength (third column); a second hypomorphic mutation has the same effect as a null mutation of any member of the pathway (last column). This phenomenon has been termed ?synergistic heterozygosity? in the context of metabolic disease 102 . � 2002 Nature Publishing Group 788 | OCTOBER 2002 | VOLUME 3 www.nature.com/reviews/genetics REVIEWS association of ABCA4 (ATP-binding cassette, subfamily A, member 4) alleles with age-related macular degenera- tion 98?101 . Although analysing the complex interactions between numerous genes and the environment remains a distant goal, modelling the interaction of a discrete number of loci and understanding the phenotypic con- sequences of such interactions is a small, yet significant step in the right direction. Perhaps the greatest opportunity stems from the promise that an expansion of monogenic concepts can assist in the modelling of multifactorial disorders, which are of great socio-economic relevance. The identification of rare, monogenic or oligogenic phenotypes that are present in complex disease has already led to the identifi- cation of susceptibility loci for complex traits, such as the presence of some CF alleles in rhinosinusitis 27 and the 1. Scriver, C. R. & Waters, P. J. Monogenic traits are not simple: lessons from phenylketonuria. Trends Genet. 15, 267?272 (1999). 2. Dipple, K. M. & McCabe, E. R. B. Phenotypes of patients with ?simple? Mendelian disorders are complex traits: thresholds, modifiers, and systems dynamics. Am. J. Hum. Genet. 66, 1729?1735 (2000). Illustrates the increasing levels of complexity that are seen in ?simple? Mendelian disorders and presents thresholds models to correlate mutations with phenotypes. 3. Dipple, K. M. & McCabe, E. R. B. Modifier genes convert ?simple? Mendelian disorders to complex traits. Mol. Genet. Metab. 71, 43?50 (2000). 4. Weiss, K. M. Is there a paradigm shift in genetics? Lessons from the study of human diseases. Mol. Phylogenet. Evol. 5, 259?265 (1996). 5. Jervis, G. A. Phenylpyruvic oligophrenia deficiency of phenylalanine-oxidizing system. Proc. Soc. Exp. Biol. Med. 82, 514?515 (1953). 6. Guthrie, R. & Susi, A. A simple phenylalanine method for detecting phenylketonuria in large populations of newborn infants. Pediatrics 32, 338?343 (1963). 7. Guthrie, R. The introduction of newborn screening for phenylketonuria: a personal history. Eur. J. Pediat. 155, 4?5 (1996). 8. Scriver, C. R. Whatever happened to PKU? Clin. Biochem. 28, 137?144 (1995). Comprehensive revision of phenylketonuria and its history, as well as a useful discussion of the complex genetics of this disorder. 9. Woo, S. L. C., Lidsky, A. S., Guttler, F., Chandra, D. & Robson, K. J. H. Cloned human phenylalanine hydroxylase gene allows prenatal diagnosis and carrier detection of classical phenylketonuria. Nature 306, 151?155 (1983). 10. Blau, N., Thony, B., Heizmann, C. W. & Dhondt, J.-L. Tetrahydrobiopterin deficiency: from phenotype to genotype. Pteridines 4, 1?10 (1993). 11. Enns, G. M. et al. Molecular correlations in phenylketonuria: mutation patterns and corresponding biochemical and clinical phenotypes in a heterogeneous California population. Pedriatr. Res. 46, 594?602 (1999). 12. Tsui, L. C. et al. Cystic fibrosis locus defined by a genetically linked polymorphic DNA marker. Science 230, 1054?1057 (1985). 13. Riordan, J. R. et al. Identification of the cystic fibrosis gene: cloning and characterization of complementary DNA. Science 245, 1066?1073 (1989). 14. Acton, J. D. & Wilmott, R. W. Phenotype of CF and the effects of possible modifier genes. Pediatr. Respir. Rev. 2, 332?339 (2001). 15. Drumm, M. L. Modifier genes and variation in cystic fibrosis. Resp. Res. 2, 125?128 (2001). 16. Mickle, J. E. & Cutting, G. R. Genotype?phenotype relationships in cystic fibrosis. Med. Clin. N. Am. 84, 597?607 (2000). 17. Nadeau, J. H. Modifier genes in mice and humans. Nature Rev. Genet. 2, 165?174 (2001). An elegant and comprehensive discussion of modifier genes and the lessons that can be learned from studying mouse models of human disease. 18. Zielenski, J. Genotype and phenotype in cystic fibrosis. Respiration 67, 117?133 (2000). 19. Zielenski, J. et al. Detection of a cystic fibrosis modifier locus for meconium ileus on human chromosome 19q13. Nature Genet. 22, 128?129 (1999). 20. Rozmahel, R. et al. Modulation of disease severity in cystic fibrosis transmembrane conductance regulator deficient mice by a secondary genetic factor. Nature Genet. 12, 280?287 (1996). 21. Arkwright, P. D. et al. TGF-?(1) genotype and accelerated decline in lung function of patients with cystic fibrosis. Thorax 55, 459?462 (2000). 22. Hull, J. & Thomson, A. H. Contribution of genetic factors other than CFTR to disease severity in cystic fibrosis. Thorax 53, 1018?1021 (1998). 23. Aron, Y. et al. HLA class II polymorphism in cystic fibrosis. A possible modifier of pulmonary phenotype. Am. J. Respir. Crit. Care Med. 159, 1464?1468 (1999). 24. Garred, P. et al. Association of mannose-binding lectin gene heterogeneity with severity of lung disease and survival in cystic fibrosis. J. Clin. Invest. 104, 431?437 (1999). 25. Grasemann, H. et al. Airway nitric oxide levels in cystic fibrosis patients are related to a polymorphism in the neuronal nitric oxide synthase gene. Am. J. Respir. Crit. Care Med. 162, 2172?2176 (2000). 26. Parmley, R. R. & Gendler, S. J. Cystic fibrosis mice lacking Muc1 have reduced amounts of intestinal mucus. J. Clin. Invest. 102, 1798?1806 (1998). 27. Wang, X. et al. Mutation in the gene responsible for cystic fibrosis and predisposition to chronic rhinosinusitis in the general population. JAMA 284, 1814?1819 (2000). 28. Groman, J. D., Meyer, M. E., Wilmott, R. W., Zeitlin, P. L. & Cutting, G. R. Variant cystic fibrosis phenotypes in the absence of CFTR mutations. N. Engl. J. Med. 347, 401?407 (2002). 29. Polanyi, M. Life?s irreducible structure. Live mechanisms and information in DNA are boundary conditions with a sequence of boundaries above them. Science 160, 53?91 (1968). 30. Strohman, R. Maneuvering in the complex path from genotype to phenotype. Science 296, 701?703 (2002). 31. Sawa, A. & Snyder, S. H. Schizophrenia: diverse approaches to a complex disease. Science 296, 692?695 (2002). An in-depth review of approaches to studying the pathogenesis of schizophrenia. 32. Hand, C. K. & Rouleau, G. A. Familial amyotrophic lateral scloerosis. Muscle Nerve 25, 135?159 (2002). 33. Giess, R. et al. Early onset of severe familial amyotrophic lateral sclerosis with a SOD-1 mutation: potential impact of CNTF as a candidate modifier gene. Am. J. Hum. Genet. 70, 1277?1286 (2002). Description of how, in an ALS patient, CNTF modulates the severity of a SOD1 mutation. 34. Takahashi, R. et al. A null mutation in the human CNTF gene is not causally related to neurological diseases. Nature Genet. 7, 79?84 (1994). 35. Su, L. K. et al. A germ line mutation of the murine homolog of the APC gene causes multiple intestinal neoplasia. Science 256, 668?670 (1992). 36. Moser, A. R., Dove, W. F., Roth, K. A. & Gordon, J. I. The Min (multiple intestinal neoplasia) mutation: its effect on gut epithelial cell differentiation and interaction with a modifier system. J. Cell Biol. 116, 1517?1526 (1992). 37. Moser, A. R., Pitot, H. C. & Dove, W. F. A dominant mutation that predisposes to multiple intestinal neoplasia in the mouse. Science 247, 322?324 (1990). 38. Dietrich, W. F. et al. Genetic identification of Mom-1, a major modifier locus affecting Min-induced intestinal neoplasia in the mouse. Cell 75, 631?639 (1993). 39. Coleman, D. L. & Eicher, E. M. Fat (fat) and Tubby (tub): two autosomal recessive mutations causing obesity syndromes in the mouse. J. Hered. 81, 424?427 (1990). 40. Noben-Trauth, K., Naggert, J. K., North, M. A. & Nishina, P. M. A candidate gene for the mouse mutation tubby. Nature 380, 534?538 (1996). 41. Kleyn, P. W. et al. Identification and characterization of the mouse obesity gene tubby: a member of a novel gene family. Cell 85, 281?290 (1996). 42. Ikeda, A. et al. Genetic modification of hearing in tubby mice: evidence for the existence of a major gene (Moth1) which protects tubby mice from hearing loss. Hum. Mol. Genet. 8, 1761?1767 (1999). 43. Ikeda, A. et al. Microtubule-associated protein 1A is a modifier of tubby hearing (moth1). Nature Genet. 30, 401?405 (2002). Cloning of the genetic modifier of tubby and the demonstration, through the use of transgenic animals, that sequence polymorphisms in Map1a are responsible for protection against the hearing defect of tubby mutant mice. 44. Kajiwara, K., Berson, E. L. & Dryja, T. P. Digenic retinitis pigmentosa due to mutations at the unlinked peripherin/RDS and ROM1 loci. Science 264, 1604?1608 (1994). Shows that heterozygous mutations in two distinct genes are necessary and sufficient to cause RP in some patients. Probably the first example of digenic mutations in humans. 45. Rivolta, C., Sharon, D., DeAngelis, M. M. & Dryja, T. P. Retinitis pigmentosa and allied diseases: numerous diseases, genes, and inheritance patterns. Hum. Mol. Genet. 11, 1219?1227 (2002). 46. Katsanis, N. et al. Mutations in MKKS cause obesity, retinal dystrophy and renal malformations associated with Bardet?Biedl syndrome. Nature Genet. 26, 67?70 (2000). 47. Slavotinek, A. M. et al. Mutations in MKKS cause Bardet?Biedl syndrome. Nature Genet. 26, 15?16 (2000). 48. Beales, P. L. et al. Genetic and mutational analyses of a large multiethnic Bardet?Biedl cohort reveal a minor involvement of BBS6 and delineate the critical intervals of other loci. Am. J. Hum. Genet. 68, 606?616 (2001). 49. Nishimura, D. Y. et al. Positional cloning of a novel gene on chromosome 16q causing Bardet?Biedl syndrome (BBS2). Hum. Mol. Genet. 10, 865?874 (2001). 50. Katsanis, N. et al. Triallelic inheritance in Bardet?Biedl syndrome, a Mendelian recessive disorder. Science 293, 2256?2259 (2001). First example of triallelic inheritance in humans, in a disease considered historically to be an autosomal-recessive trait. 51. Feingold, E. Regression-based quantitative-trait-Locus mapping in the 21st century. Am. J. Hum. Genet. 71, 217?222 (2002). 52. Moore, J. H. & Williams, S. M. New strategies for identifying gene-gene interactions in hypertension. Ann. Med. 34, 88?95 (2002). 53. McCallion, A. S. & Chakravarti, A. in Inborn Errors of Development (eds Epstein, C., Erickson, R. & Wynshaw- Boris, A. (Oxford Univ. Press, San Francisco, in the press). 54. Puffenberger, E. G. et al. A missense mutation of the endothelin-B receptor gene in multigenic Hirschsprung?s disease. Cell 79, 1257?1266 (1994). 55. Parisi, M. A. & Kapur, R. P. Genetics of Hirschsprung disease. Curr. Opin. Pediatr. 12, 610?617 (2000). 56. Amiel, J. & Lyonett, S. Hirschsprung disease, associated syndromes, and genetics: a review. J. Med. Genet. 38, 729?739 (2001). 57. Angrist, M. et al. A gene for Hirschsprung disease (megacolon) in the pericentromeric region of human chromosome 10. Nature Genet. 3, 351?356 (1993). 58. Lyonnet, S. et al. A gene for Hirschsprung disease maps to the proximal long arm of chromosome 10. Nature Genet. 4, 346?350 (1993). 59. Bolk-Gabriel, S. B. et al. Segregation at three loci explains familial and population risk in Hirschsprung disease. Nature Genet. 31, 89?93 (2002). 60. Bolk, S. et al. A human model for multigenic inheritance: phenotypic expression in Hirschsprung disease requires both the RET gene and a new 9q31 locus. Proc. Natl Acad. Sci. USA 97, 268?273 (2000). References 59 and 60 establish statistical methods to analyse the documented oligogenicity of Hirschsprung disease. 61. Floeth, M. & Bruckner-Tuderman, L. Digenic junctional epidermolysis bullosa: mutations in COL17A1 and LAMB3 genes. Am. J. Hum. Genet. 65, 1530?1537 (1999). � 2002 Nature Publishing Group NATURE REVIEWS | GENETICS VOLUME 3 | OCTOBER 2002 | 789 REVIEWS 62. Kestila, M. et al. Positionally cloned gene for a novel glomerular protein nephrin is mutated in congenital nephrotic syndrome. Mol. Cell 1, 575?582 (1998). 63. Patrakka, J. et al. Congenital nephrotic syndrome (NPHS1): features resulting from different mutations in Finnish patients. Kidney Int. 58, 972?980 (2000). 64. Boute, N. et al. NPHS2, encoding the glomerular protein podocin, is mutated in autosomal recessive steroid-resistant nephrotic syndrome. Nature Genet. 24, 349?354 (2000). 65. Huber, T. B., Kottgen, M., Schilling, B., Walz, G. & Benzing, T. Interaction with podocin facilitates nephrin signaling. J. Biol. Chem. 276, 1543?1546 (2001). 66. Koziell, A. et al. Genotype/phenotype correlations of NPHS1 and NPHS2 mutations in nephrotic syndrome advocate a functional inter-relationship in glomerular filtration. Hum. Mol. Genet. 11, 379?388 (2002). 67. Mykytyn, K. et al. Identification of the gene that, when mutated, causes the human obesity syndrome BBS4. Nature Genet. 28, 188?191 (2001). 68. Katsanis, N. et al. BBS4 is a minor contributor to Bardet?Biedl syndrome and may also participate in triallelic inheritance. Am. J. Hum. Genet. 71, 22?29 (2002). 69. Stone, E. M. et al. Identification of a gene that causes primary open angle glaucoma. Science 275, 668?670 (1997). 70. Stoilov, I., Akarsu, A. N. & Sarfarazi, M. Identification of three different truncating mutations in cytochrome P450B1 (CYP1B1) as the principal cause of primary congenital glaucoma (buphthalmos) in families linked to the GLC3A locus on chromosome 2p21. Hum. Mol. Genet. 6, 641?647 (1997). 71. Fingert, J. H. et al. Analysis of myocilin mutations in 1703 glaucoma patients from five different populations. Hum. Mol. Genet. 8, 899?905 (1999). 72. Bejjani, B. A. et al. Multiple CYP1B1 mutations and incomplete penetrance in an inbred population segregating primary congenital glaucoma suggest frequent de novo events and a dominant modifier locus. Hum. Mol. Genet. 12, 367?374 (2000). 73. Bejjani, B. A. et al. Mutations in CYP1B1, the gene for cytochrome P4501B1, are the predominant cause of primary congenital glaucoma in Saudi Arabia. Am. J. Hum. Genet. 62, 325?333 (1998). 74. Vincent, A. L. et al. Digenic inheritance of early-onset glaucoma: CYP1B1, a potential modifier gene. Am. J. Hum. Genet. 70, 448?460 (2002). 75. Hanna, I. H., Dawling, S., Roodi, N., Guengerich, F. P. & Parl, F. F. Cytochrome P450 1B1 (CYP1B1) pharmacogenetics: association of polymorphisms with functional differences in hydrogen hydroxylation activity. Cancer Res. 60, 3440?3444 (2000). 76. Shimada, T., Watanabe, J., Inoue, K., Guengerich, F. P. & Gillam, E. M. Specificity of 17b-oestradiol and benzo[a]pyrene oxidation by polymorphic human cytochrome P450B1 variants substituted at residues 48, 119 and 432. Xenobiotica 31, 163?176 (2001). 77. Shimada, T. et al. Catalytic properties of polymorphic human cytochrome P450 1B1 variants. Carcinogenesis 20, 1607?1613 (1999). 78. Li, D. N., Seidel, A., Pritchard, M. P., Wolf, C. R. & Friedberg, T. Polymorphisms in P450 CYP1B1 affect the conversion of estradiol to the potentially carcinogenic metabolite 4- hydroxyestradiol. Pharmacogenetics 10, 343?353 (2000). 79. Goldberg, A. F. X. & Molday, R. S. Subunit composition of the peripherin/rds-rom-1 disk rim complex from rod photoreceptors: hydrodynamic evidence for a tetrameric quaternary structure. Biochemistry 35, 6144?6149 (1996). 80. Clark, G. et al. Rom-1 is required for rod photoreceptor viability and the regulation of disk morphogenesis. Nature Genet. 25, 67?73 (2000). 81. Travis, G. H., Brennan, M. B., Danielson, P. E., Kozak, C. A. & Sutcliffe, J. G. Identification of a photoreceptor-specific mRNA encoded by the gene responsible for retinal degeneration slow (rds). Nature 338, 70?73 (1989). 82. Loewen, C. J., Moritz, O. L. & Molday, R. S. Molecular characterization of the peripherin-2 and rom-1 mutants responsible for digenic retinitis pigmentosa. J. Biol. Chem. 276, 22388?22396 (2001). 83. Goldberg, A. F. & Molday, R. S. Defective subunit assembly underlies a digenic form of retinitis pigmentosa linked to mutations in peripherin/rds and rom-1. Proc. Natl Acad. Sci. USA 93, 13726?13730 (1996). References 82 and 83 illustrate the molecular basis of digenic inheritance involving ROM1 and RDS in human RP. 84. Schwartz, K. et al. Podocin, a raft-associated component of the glomerular slit diaphragm, interacts with CD2AP and nephrin. J. Clin. Invest. 108, 1621?1629 (2001). 85. Angrist, M., Bolk, S., Halushka, M., Lapchak, P. & Chakravarti, A. Germline mutations in glial cell-derived neurotrophic factor (GDNF) and RET in a Hirschsprung disease patient. Nature Genet. 14, 341?344 (1996). 86. Durbec, P. et al. GDNF signaling through the Ret receptor tyrosine kinase. Nature 381, 789?793 (1996). 87. Treanor, J. J. et al. Characterization of a multicomponent receptor for GDNF. Nature 382, 80?83 (1996). 88. Li, L. et al. Alagille syndrome is caused by mutations in human Jagged1, which encodes a ligand for Notch1. Nature Genet. 16, 243?251 (1997). 89. Oda, T. et al. Mutations in the human Jagged1 gene are responsible for Alagille syndrome. Nature Genet. 16, 235?242 (1997). 90. McCright, B., Lozier, J. & Gridley, T. A mouse model for Alagille syndrome: Notch2 as a genetic modifier of Jag1 haploinsufficiency. Development 129, 1075?1082 (2002). Describes the molecular basis of phenotypic modification of the Jag1 mutation. 91. Yook, K. J., Proulx, S. R. & Jorgensen, E. M. Rules of nonallelic complementation at the synapse in Caenorhabditis elegans. Genetics 158, 209?220 (2001). A comprehensive analysis of non-allelic non- complementation and establishment of its parameters of action in the worm synapse. 92. Fuller, M. T. et al. Interacting genes identify interacting proteins involved in microtubule function in Drosophila. Cell Motil. Cytoskeleton 14, 128?135 (1989). 93. Stearns, T. & Botstein, D. Unlinked noncomplementation: isolation of new conditional-lethal mutations in each of the tubulin genes of Saccharomyces cerevisiae. Genetics 119, 249?260 (1988). A classic paper on the use of non-allelic non- complementation to study specific signalling pathways. 94. Westphal, l. V. et al. A frequent mild mutation in ALG6 may exacerbate the clinical severity of patients with congenital disorder of glycosylation Ia (CDG-Ia) caused by phosphomannomutase deficiency. Hum. Mol. Genet. 11, 599?604 (2002). 95. Gavin, A. C. et al. Functional organization of the yeast proteome by systematic analysis of protein complexes. Nature 415,41?47 (2002). 96. Ho, Y. et al. Systematic identification of protein complexes in Saccharomyces cerevisiae by mass spectrometry. Nature 415, 180?183 (2002). 97. Scriver, C. R. Why mutation analysis does not always predict clinical consequences: explanations in the era of genomics. J. Pediatr. 140, 502?506 (2002). 98. Allikmets, R. et al. Mutation of the Stargardt disease gene (ABCR) in age-related macular degeneration. Science 277, 1805?1807 (1997). 99. Allikmets, R. et al. A photoreceptor cell-specific ATP-binding transporter gene (ABCR) is mutated in recessive Stargardt macular dystrophy. Nature Genet. 15, 236?246 (1997). 100. Sun, H., Smallwood, P. M. & Nathans, J. Biochemical defects in ABCR protein variants associated with human retinopathies. Nature Genet. 26, 242?246 (2000). 101. Shroyer, N. F., Lewis, R. A., Yatsenko, A. N., Wensel, T. G. & Lupski, J. R. Cosegregation and functional analysis of mutant ABCR (ABCA4) alleles in families that manifest both Stargardt disease and age-related macular degeneration. Hum. Mol. Genet. 10, 2671?2678 (2001). References 100 and 101 show the mutagenic potential of alleles for which pathogenicity had been disputed according to genetic criteria. 102. del Castillo, I. et al. A deletion involving the connexin 30 gene in nonsyndromic hearing impairment. N. Engl. J. Med. 346, 243?249 (2002). 103. Riazuddin, S. et al. Dominant modifier DFNM1 suppresses recessive deafness DFNB26. Nature Genet. 26, 431?434 (2000). 104. Adato, A. et al. Possible interaction between USH1B and USH3 gene products as implied by apparent digenic deafness inheritance. Am. J. Hum. Genet. 65, 261?265 (1999). 105. Balciuniene, J. et al. Evidence for digenic inheritance of nonsyndromic hereditary hearing loss in a Swedish family. Am. J. Hum. Genet. 63, 786?793 (1998). 106. Morell, R. et al. Apparent digenic inheritance of Waardenburg syndrome type 2 (WS2) and autosomal recessive ocular albinism (AROA). Hum. Mol. Genet. 6, 659?664 (1997). 107. Richard, I. et al. Mutations in the proteolytic enzyme Calpain 3 cause limb-girdle muscular dystrophy type 2A. Cell 81, 27?40 (1995). 108. Bolliger-Stucki, B., Lord, S. T. & Furlan, M. Fibrinogen Milano XII: a dysfunctional variant containing 2 amino acid substitutions, ?-R16C and ?-G165R. Blood 98, 351?357 (2001). 109. Goodyer, P. R., Clow, C., Reade, T. & Girardin, C. Prospective analysis and classification of patients with cystinuria identified in a newborn screening program. J. Pediatr. 122, 568?572 (1993). 110. Kerst, B. et al. Heterozygous myogenic factor 6 mutation associated with myopathy and severe course of Becker muscular dystrophy. Neuromuscul. Disord. 10, 572?577 (2000). 111. Redston, M. et al. The APCI1307K allele and breast cancer risk. Nature Genet. 20, 13?14 (1998). 112. Scharf, J. M. et al. Identification of a candidate modifying gene for spinal muscular atrophy by comparative genomics. Nature Genet. 20, 83?86 (1998). 113. Phelan, C. M. et al. Ovarian cancer risk in BRCA1 carriers is modified by the HRAS1 variable number of tandem repeat (VNTR) locus. Nature Genet. 12, 309?311 (1996). 114. Knoblauch, H. et al. A cholesterol-lowering gene maps to chromosome 13q. Am. J. Hum. Genet. 66, 157?166 (2000). 115. Cazeneuve, C. et al. Identification of MEFV-independent modifying genetic factors for familial Mediterranean fever. Am. J. Hum. Genet. 67, 1136?1143 (2000). 116. Bykhovskaya, Y. et al. Candidate locus for a nuclear modifier gene for maternally inherited deafness. Am. J. Hum. Genet. 66, 1905?1910 (2000). 117. Box, N. F. et al. MC1R genotype modifies risk of melanoma in families segregating CDKN2A mutations. Am. J. Hum. Genet. 69, 765?773 (2001). 118. Sertie, A. L., Sousa, A. V., Steman, S., Pavanello, R. C. & Passos-Bueno, M. R. Linkage analysis in a large Brazilian family with van der Woude syndrome suggests the existence of a susceptibility locus for cleft palate at 17p11. ?2-11.?1. Am. J. Hum. Genet. 65, 433?440 (1999). 119. Gill, J. C., Endres-Brooks, J., Bauer, P. J., Marks, W. J. J. & Montgomery, R. R. The effect of ABO blood group on the diagnosis of von Willebrand disease. Blood 69, 1691?1695 (1987). 120. Mohlke, K. L. et al. Mvwf, a dominant modifier of murine von Willebrand factor, results from altered lineage-specific expression of a glycosyltransferase. Cell 96, 111?120 (1999). 121. Wyss-Coray, T. et al. TGF-?1 promotes microglial amyloid-? clearance and reduces plaque burden in transgenic mice. Nature Med. 7, 612?618 (2001). 122. Takaku, K. et al. Suppression of intestinal polyposis in Apc ?716 knockout mice by an additional mutation in the cytosolic phospholipase A 2 gene. J. Biol. Chem. 275, 34013?34016 (2000). 123. Oshima, M. et al. Suppression of intestinal polyposis in Apc ?716 knockout mice by inhibition of cyclooxygenase 2 (COX-2). Cell 87, 803?809 (1996). 124. Vockley, J., Rinaldo, P., Bennett, M. J., Matern, D. & Vladutiu, G. D. Synergistic heterozygosity: disease resulting from multiple partial defects in one or more metabolic pathways. Mol. Genet. Metab. 71, 10?18 (2000). Acknowledgements We apologize to those researchers whose work we were unable to represent due to space constraints. We thank J. Groman and G. Cutting for sharing their unpublished data with us, and G. Cutting, S. Huston and J. Lupski for constructive discussions and critique of this manuscript. This work was supported by the National Eye Institute, the National Institutes of Health and by a grant from March of Dimes to N.K. Online links DATABASES The following terms in this article are linked online to: LocusLink: http://www.ncbi.nlm.nih.gov/LocusLink ALG6 | APC | Apc | BBS2 | BBS6 | CFTR | COL17A1 | CNTF | CYP1B1 | EDNRB | FAP| GDNF | HLA class II | JAG1 | LAMB3 | Map1a | MBL | Mom1 | moth1 | MTAP1 | Muc1 | NOS1 | NOTCH2 | NPHS1 | NPHS2 | PAH | PMM2 | RDS | ROM1 | SOD1 | TGFB1 | tub OMIM: http://www.ncbi.nlm.nih.gov/Omim Bardet?Biedl syndrome | cystic fibrosis | familial amyotrophic lateral sclerosis | FAP | Finnish congenital nephrotic syndrome | generalized atrophic benign epidermolysis bullosa | Hirschsprung disease | hyperphenylalaninaemia | junctional epidermolysis bullosa | juvenile open-angle glaucoma | PKU | RP |schizophrenia | segmental glomerulosclerosis WormBase: http://wormbase.org UNC-13 FURTHER INFORMATION Nicholas Katsanis? lab: http://www.hopkinsmedicine.org/geneticmedicine/Faculty/ FacultyProfile.cfm?ProfileID=10 Access to this interactive links box is free online. "
Add Content to Group
|
Bookmark
|
Keywords
|
Flag Inappropriate
share
Close
Digg
Facebook
MySpace
Google+
Comments
Close
Please Post Your Comment
*
The Comment you have entered exceeds the maximum length.
Submit
|
Cancel
*
Required
Comments
Please Post Your Comment
No comments yet.
Save Note
Note
View
Public
Private
Friends & Groups
Friends
Groups
Save
|
Cancel
|
Delete
Please provide your notes.
Next
|
Prev
|
Close
|
Edit
|
Delete
Genetics
Gene Inheritance and Transmission
Gene Expression and Regulation
Nucleic Acid Structure and Function
Chromosomes and Cytogenetics
Evolutionary Genetics
Population and Quantitative Genetics
Genomics
Genes and Disease
Genetics and Society
Cell Biology
Cell Origins and Metabolism
Proteins and Gene Expression
Subcellular Compartments
Cell Communication
Cell Cycle and Cell Division
Scientific Communication
Career Planning
Loading ...
Scitable Chat
Register
|
Sign In
Visual Browse
Close
Comments
CloseComments
Please Post Your Comment