Fishing for the ‘Missing Heritability’ of ALS

Scientists have not nearly plumbed the depths of genetic associations with sporadic amyotrophic lateral sclerosis, according to a new analysis in the July 14 JAMA Neurology online. First author Margaux Keller and colleagues at the National Institute on Aging in Bethesda, Maryland, measured heritability—the degree to which ALS risk factors are inherited, as opposed to attributable to one’s environment. Combining data from three separate genome-wide association studies (GWAS), the authors conclude that if one could sum up all the factors that led to ALS in the combined 1,223 cases studied, genetic heritability would account for at least 21 percent of the risk. Only 0.5 percent of ALS risk comes from known ALS genes, Keller estimated; the rest are unknown. The authors identified 11 new chromosomal regions that might harbor some of the missing heritability—prime areas to trawl for the next round of ALS genes.

The study backs up a recent trend in thinking among ALS geneticists: that people who have apparently sporadic disease can nonetheless blame their genes for their condition, at least in part, said Summer Gibson of the University of Utah School of Medicine in Salt Lake City. (See also Jan 2014 news story.) Gibson was not involved in the study. She said the 21 percent figure is not one she would quote to patients and their relatives; it does not mean that for an individual patient, genes explain 21 percent of their disease. Rather, it applies to a population of sporadic ALS cases, in which overall about a fifth of disease risk came from genes. Thus, the paper tells geneticists that they should keep looking. 

New Method Gauges Heritability
To quantify this ALS heritability, Keller and senior authors Bryan Traynor and Michael Nalls decided to apply a relatively new kind of analysis called genome-wide complex trait analysis (GCTA) (Yang et al., 2013; Yang et al., 2011). GCTA differs from a typical GWAS, which analyzes each single-nucleotide polymorphism (SNP) individually, asking if the particular variant contributes to disease risk. A GCTA considers all SNPs at once, asking how much of this overall genetic variety contributes to disease risk. This is an advantage, Keller said, because it allows the overall effect of multiple SNPs, which are non-significant on their own but are somehow additive or multiplicative, to be measured.

A disadvantage, Keller added, is that only SNPs are analyzed; de novo mutations would be missed, as would genetic lesions such as insertions or deletions. In ALS, a large series of repeats in the C9ORF72 gene causes a large number of familial cases (see Sep 2011 news story); similar genetic defects would be invisible to a GCTA study. In Keller’s analysis, the C9ORF72 region did not make a major contribution to heritability, though nearby SNPs did; these could be variants that co-segregate with the repeat region, Keller said, but she could not be sure.

Keller and colleagues have already applied GCTA to Parkinson’s disease (Keller et al., 2012), and another group recently applied it to ALS as well (Fogh et al., 2014). These latter researchers attributed 12 percent of ALS to genetics. The difference highlights some of the factors to which GCTA results are sensitive. “GCTA is a well-established technique, but the results can be affected considerably by the estimate of disease prevalence used,” commented Ammar Al-Chalabi of King’s College London, a co-author of the ALS GCTA paper earlier this year, in an email to Alzforum (see full comment below). Keller and colleagues used a prevalence of one in 10,000, whereas the other study authors used 0.5 in 10,000. In the earlier study, Al-Chalabi and colleagues also calculated heritability for a one in 10,000 disease prevalence, and then they got the same number as Keller. “It could be argued that 12 to 21 percent is a large difference, but in absolute terms it is only 9 percent, which is not that huge,” Al-Chalabi noted.

Heritability also varies by population. The Keller study included cohorts from Finland (Laaksovirta et al., 2010), the United States (see Feb 2009 news story; Schymick et al., 2007), and Italy (Traynor et al., 2010). The other ALS GCTA included some of the same Italians, but also other cohorts from Europe and the United States, so the heritability figures of 12 and 21 percent apply to somewhat different populations.

The Next ALS Gene Addresses Identified?
All told, the data sets Keller used offered 1,223 ALS (both sporadic and familial) and 1,591 control SNP genotypes. She estimated that sporadic ALS, in the European-American population studied, results from 0.5 percent known genes and 20.5 percent yet-to-be-discovered genes. The 21 percent figure is a lowball estimate, the authors note, because GCTA misses insertions, deletions, and de novo mutations. The remaining 79 percent of ALS cause could include more genes, environmental factors, or gene-environment interactions, Keller said.

Keller and colleagues took their analysis a step further. “What is novel about this study is the use of GCTA to attempt to localize the genomic regions that might hold the relevant gene variants,” Al-Chalabi commented. The authors divided the genome into different segments of about 20 megabases in size and applied GCTA to each. The genome regions with high heritability figures likely contain genes that influence ALS risk, they surmised. “It does not tell us which gene within a given region is responsible, but it gives us new regions to examine in greater detail,” Keller said.

This analysis identified 17 regions of interest. Six of them already contain known ALS genes, such as the part of chromosome nine around C9ORF72 and Valosin-Containing Protein. Eleven others are not yet known to host ALS genes, and direct sequencing of those segments might identify new genes, Keller suggested. The authors were particularly intrigued by areas on chromosomes 15 and 22, but declined to speculate which genes might be responsible without biological data to back them up. “‘Missing heritability’ [is a] term that will likely disappear from the vernacular as genome sequencing data sets grow in size,” they wrote.—Amber Dance

Comments

  1. GCTA is now a well-established technique, but the result can be affected considerably by the estimate of disease prevalence used. At the prevalence used here by Margaux Keller and colleagues, the 21 percent heritability estimate is identical to the result we published earlier this year using the same method and an overlapping data set (Fogh et al., 2014). When a prevalence estimate of five per 100,000 is used however, the heritability estimate is closer to 12 percent. Without doubt there remains some genetic contribution to ALS that remains unidentified, and this is true for apparently sporadic as well as familial ALS. For example, it is easy to show that every familial disease gene should be found in apparently sporadic cases and that the distinction between familial and sporadic disease is not clear-cut (Al-Chalabi et al., 2011; Byrne et al., 2012; Hanby et al., 2011). What is novel about this study is the use of GCTA to attempt to localize the genomic regions that might hold the relevant gene variants, and it will be interesting to see if the predictions hold up.

    It is likely that future ALS gene discoveries will include some with high and some with modest penetrance, and the contribution to risk in any individual is probably a combination of multiple gene variants and other risk exposures (Al-Chalabi et al., 2013). We already have examples of gene variants of low penetrance in UNC13A and ANG, moderate penetrance in C9orf72 and high penetrance in SOD1, TARDBP, FUS etc., and this trend will continue. A combination of patient engagement in research, large research collaborations, stupendous advances in gene technology and the availability of supercomputing make the discovery of the complete genetic architecture of ALS a realistic possibility.

    References:

    Fogh I, Ratti A, Gellera C, Lin K, Tiloca C, Moskvina V, Corrado L, Sorarù G, Cereda C, Corti S, Gentilini D, Calini D, Castellotti B, Mazzini L, Querin G, Gagliardi S, Del Bo R, Conforti FL, Siciliano G, Inghilleri M, Saccà F, Bongioanni P, Penco S, Corbo M, Sorbi S, Filosto M, Ferlini A, Di Blasio AM, Signorini S, Shatunov A, Jones A, Shaw PJ, Morrison KE, Farmer AE, Van Damme P, Robberecht W, Chiò A, Traynor BJ, Sendtner M, Melki J, Meininger V, Hardiman O, Andersen PM, Leigh NP, Glass JD, Overste D, Diekstra FP, Veldink JH, van Es MA, Shaw CE, Weale ME, Lewis CM, Williams J, Brown RH, Landers JE, Ticozzi N, Ceroni M, Pegoraro E, Comi GP, D'Alfonso S, van den Berg LH, Taroni F, Al-Chalabi A, Powell J, Silani V, SLAGEN Consortium and Collaborators. A genome-wide association meta-analysis identifies a novel locus at 17q11.2 associated with sporadic amyotrophic lateral sclerosis. Hum Mol Genet. 2014 Apr 15;23(8):2220-31. Epub 2013 Nov 20 PubMed.

    Al-Chalabi A, Lewis CM. Modelling the effects of penetrance and family size on rates of sporadic and familial disease. Hum Hered. 2011;71(4):281-8. PubMed.

    Byrne S, Elamin M, Bede P, Hardiman O. Absence of consensus in diagnostic criteria for familial neurodegenerative diseases. J Neurol Neurosurg Psychiatry. 2012 Apr;83(4):365-7. PubMed.

    Hanby MF, Scott KM, Scotton W, Wijesekera L, Mole T, Ellis CE, Leigh PN, Shaw CE, Al-Chalabi A. The risk to relatives of patients with sporadic amyotrophic lateral sclerosis. Brain. 2011 Dec;134(Pt 12):3454-7. Epub 2011 Sep 20 PubMed.

    Al-Chalabi A, Hardiman O. The epidemiology of ALS: a conspiracy of genes, environment and time. Nat Rev Neurol. 2013 Nov;9(11):617-28. Epub 2013 Oct 15 PubMed.

    View all comments by Ammar Al-Chalabi

  2. This paper shows that sporadic ALS is indeed caused by genetic factors, as had been implied by previous twin studies. The method used in this paper shows that the contribution of genes is considerable and higher than estimated previously using GWAS data. This reinforces the need for genetic research in ALS.

    The missing heritability in ALS is likely to be in rarer genetic variants with relatively large effect. The way to find these variants is by performing even larger GWAS (see our own effort: project MINE) using better chips with higher SNPs densities and imputation, analyzing pedigrees and extended pedigrees. As genotyping methods improve and costs drop, we will likely be able to perform large whole exome-/genome-sequencing studies allowing us to find these rare variants with large effects. Luckily, there are many collaborative efforts within the field of ALS which make it possible to get these large studies done.

    View all comments by Michael van Es

Make a Comment

To make a comment you must login or register.

References

News Citations

  1. Familial ALS More Common Than Thought—Do We Need a New Definition?
  2. Corrupt Code: DNA Repeats Are Common Cause for ALS and FTD
  3. Genomewide Screen for SNPs Linked to Sporadic ALS Finds…Nothing Yet

Paper Citations

  1. Yang J, Lee SH, Goddard ME, Visscher PM. Genome-wide complex trait analysis (GCTA): methods, data analyses, and interpretations. Methods Mol Biol. 2013;1019:215-36. PubMed.
  2. Yang J, Lee SH, Goddard ME, Visscher PM. GCTA: a tool for genome-wide complex trait analysis. Am J Hum Genet. 2011 Jan 7;88(1):76-82. Epub 2010 Dec 17 PubMed.
  3. Keller MF, Saad M, Bras J, Bettella F, Nicolaou N, Simón-Sánchez J, Mittag F, Büchel F, Sharma M, Gibbs JR, Schulte C, Moskvina V, Durr A, Holmans P, Kilarski LL, Guerreiro R, Hernandez DG, Brice A, Ylikotila P, Stefánsson H, Majamaa K, Morris HR, Williams N, Gasser T, Heutink P, Wood NW, Hardy J, Martinez M, Singleton AB, Nalls MA, International Parkinson's Disease Genomics Consortium (IPDGC), Wellcome Trust Case Control Consortium 2 (WTCCC2). Using genome-wide complex trait analysis to quantify 'missing heritability' in Parkinson's disease. Hum Mol Genet. 2012 Nov 15;21(22):4996-5009. Epub 2012 Aug 13 PubMed.
  4. Fogh I, Ratti A, Gellera C, Lin K, Tiloca C, Moskvina V, Corrado L, Sorarù G, Cereda C, Corti S, Gentilini D, Calini D, Castellotti B, Mazzini L, Querin G, Gagliardi S, Del Bo R, Conforti FL, Siciliano G, Inghilleri M, Saccà F, Bongioanni P, Penco S, Corbo M, Sorbi S, Filosto M, Ferlini A, Di Blasio AM, Signorini S, Shatunov A, Jones A, Shaw PJ, Morrison KE, Farmer AE, Van Damme P, Robberecht W, Chiò A, Traynor BJ, Sendtner M, Melki J, Meininger V, Hardiman O, Andersen PM, Leigh NP, Glass JD, Overste D, Diekstra FP, Veldink JH, van Es MA, Shaw CE, Weale ME, Lewis CM, Williams J, Brown RH, Landers JE, Ticozzi N, Ceroni M, Pegoraro E, Comi GP, D'Alfonso S, van den Berg LH, Taroni F, Al-Chalabi A, Powell J, Silani V, SLAGEN Consortium and Collaborators. A genome-wide association meta-analysis identifies a novel locus at 17q11.2 associated with sporadic amyotrophic lateral sclerosis. Hum Mol Genet. 2014 Apr 15;23(8):2220-31. Epub 2013 Nov 20 PubMed.
  5. Laaksovirta H, Peuralinna T, Schymick JC, Scholz SW, Lai SL, Myllykangas L, Sulkava R, Jansson L, Hernandez DG, Gibbs JR, Nalls MA, Heckerman D, Tienari PJ, Traynor BJ. Chromosome 9p21 in amyotrophic lateral sclerosis in Finland: a genome-wide association study. Lancet Neurol. 2010 Oct;9(10):978-85. PubMed.
  6. Schymick JC, Scholz SW, Fung HC, Britton A, Arepalli S, Gibbs JR, Lombardo F, Matarin M, Kasperaviciute D, Hernandez DG, Crews C, Bruijn L, Rothstein J, Mora G, Restagno G, Chiò A, Singleton A, Hardy J, Traynor BJ. Genome-wide genotyping in amyotrophic lateral sclerosis and neurologically normal controls: first stage analysis and public release of data. Lancet Neurol. 2007 Apr;6(4):322-8. PubMed.
  7. Traynor BJ, Nalls M, Lai SL, Gibbs RJ, Schymick JC, Arepalli S, Hernandez D, van der Brug MP, Johnson JO, Dillman A, Cookson M, Moglia C, Calvo A, Restagno G, Mora G, Chiò A. Kinesin-associated protein 3 (KIFAP3) has no effect on survival in a population-based cohort of ALS patients. Proc Natl Acad Sci U S A. 2010 Jul 6;107(27):12335-8. PubMed.

External Citations

  1. C9ORF72
  2. Valosin-Containing Protein

Further Reading

Papers

  1. Al-Chalabi A, Lewis CM. Modelling the effects of penetrance and family size on rates of sporadic and familial disease. Hum Hered. 2011;71(4):281-8. PubMed.
  2. Byrne S, Elamin M, Bede P, Hardiman O. Absence of consensus in diagnostic criteria for familial neurodegenerative diseases. J Neurol Neurosurg Psychiatry. 2012 Apr;83(4):365-7. PubMed.
  3. Hanby MF, Scott KM, Scotton W, Wijesekera L, Mole T, Ellis CE, Leigh PN, Shaw CE, Al-Chalabi A. The risk to relatives of patients with sporadic amyotrophic lateral sclerosis. Brain. 2011 Dec;134(Pt 12):3454-7. Epub 2011 Sep 20 PubMed.
  4. Al-Chalabi A, Hardiman O. The epidemiology of ALS: a conspiracy of genes, environment and time. Nat Rev Neurol. 2013 Nov;9(11):617-28. Epub 2013 Oct 15 PubMed.
  5. Al-Chalabi A, Fang F, Hanby MF, Leigh PN, Shaw CE, Ye W, Rijsdijk F. An estimate of amyotrophic lateral sclerosis heritability using twin data. J Neurol Neurosurg Psychiatry. 2010 Dec;81(12):1324-6. Epub 2010 Sep 22 PubMed.
  6. Fang F, Kamel F, Lichtenstein P, Bellocco R, Sparén P, Sandler DP, Ye W. Familial aggregation of amyotrophic lateral sclerosis. Ann Neurol. 2009 Jul;66(1):94-9. PubMed.
  7. Manolio TA, Collins FS, Cox NJ, Goldstein DB, Hindorff LA, Hunter DJ, McCarthy MI, Ramos EM, Cardon LR, Chakravarti A, Cho JH, Guttmacher AE, Kong A, Kruglyak L, Mardis E, Rotimi CN, Slatkin M, Valle D, Whittemore AS, Boehnke M, Clark AG, Eichler EE, Gibson G, Haines JL, Mackay TF, McCarroll SA, Visscher PM. Finding the missing heritability of complex diseases. Nature. 2009 Oct 8;461(7265):747-53. PubMed.

Primary Papers

  1. Keller MF, Ferrucci L, Singleton AB, Tienari PJ, Laaksovirta H, Restagno G, Chiò A, Traynor BJ, Nalls MA. Genome-wide analysis of the heritability of amyotrophic lateral sclerosis. JAMA Neurol. 2014 Sep;71(9):1123-34. PubMed.

Annotate

To make an annotation you must Login or Register.