CoChat
Sign Up
Cited by 173Autism Spectrum Disorder ResearchGenomics and Rare DiseasesGenetics and Neurodevelopmental Disorders

Low load for disruptive mutations in autism genes and their biased transmission

Open access
Iossifov, Ivan, Levy, Dan, Allen, Jeremy, Ye, Kenny, Ronemus, Michael, Lee, Yoon-ha, Yamrom, Boris, Wigler, Michael·2015·Proceedings of the National Academy of Sciences
DOI 10.1073/pnas.1516376112PMC PMC4611648
Open this paper in CoChat — free accountOpen PDF

Abstract

Significance: Gene targets of de novo mutation in autistic children have a lighter load of rare disruptive variation than typical human genes. This finding suggests such mutations are under negative selection and autism genes are highly vulnerable to mutation. Disruptive variants in these genes have biased transmission: They are more frequently transmitted to affected children, and more often from mothers than from fathers. Targets of mutation in lower intelligence quotient (IQ) affected children have a lower load of disruptive mutations than targets of mutation in higher IQ affected children. Biased transmission is seen more frequently to affected children of lower IQ. These observations are consistent with a correlation between severity of mutations and phenotype, and based on them, we list candidate autism genes ordered by likelihood.

Integrity signals

Registry

Crossref · 10.1073/pnas.1516376112

Cites

21 of 21 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 26 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    De novo target gene sets from lower-IQ affected children have significantly lower LGD ratios than sibling targets (P = 0.0011).

    “Judging by this measure, DN target gene sets from affected children with lower IQ have significantly lower LGD ratios than DN targets from siblings (P = 0.0011).”
    Quote in the paperSubstantial claim
  • Findingown data

    After posterior probabilities including transmission adjustment, scores of target genes in various classes are bimodal.

    “After posterior probabilities are calculated, including adjustment for transmission, the scores of target genes in the various classes become bimodal (Fig. 2, Right).”
    • Also stated in: § Discussion · boosted
      “Moreover, we show a bimodal distribution in mutational load in the proportion predicted by theory for candidate autism genes and bystander target genes.”
    Quote in the paperSubstantial claim
    • The authors state this as an assumption: Autism vulnerable genes and intolerant (protected) genes are treated as the same set.
    • The authors state this as an assumption: DN and rare parental data are independent.
    • The authors state this as an assumption: A protected gene has only 10% of typical share of UR LGD mutations.
  • Findingown data

    Ultrarare missense accumulation is proportional to ultrarare synonymous accumulation.

    “By contrast, UR missense accumulation proportionately follows UR synonymous accumulation (Fig. 1B).”
    Quote in the paperSubstantial claim
  • Findingnone given

    No ascertainment differential is observed for non-disruptive mutations such as synonymous variants.

    “An ascertainment differential is not seen for mutations that are not disruptive, such as synonymous variants.”
    Quote in the paperSubstantial claim
  • Findingown data

    Ultrarare LGD variant load is not linear with ultrarare synonymous variant load.

    “The load of UR LGD variants is not linear with the load of UR synonymous variants.”
    Quote in the paperSubstantial claim
  • May combine two claims. 28
  • Classified as probably not a substantial claim. 13

  • May combine two claims.

    Findingown data

    By affected IQ, the transmission differential is 70 from lower-IQ families and 30 from higher-IQ families (P = 0.0071 and P = 0.1489).

    “Almost as striking is the decomposition by affected IQ: A differential of 70 comes from precisely half of the families with an affected child of lower IQ, and 30 come from families with an affected child of higher IQ (P = 0.0071 and P = 0.1489, respectively).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The transmission ascertainment differential for genes with one LGD variant is 101, which may explain about 5.4% of SSC families.

    “The ascertainment differential of transmission of UR LGD variants in the genes with a single LGD variant is 101, contributing to diagnosis in perhaps 5.4% of families from the SSC.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Embryonic category shows differential transmission of UR LGD variants to affected children over siblings; 31 more transmitted only to affected child among 298 opportunities (P=0.0073).

    “The embryonically expressed category shows a surprising differential transmission of UR LGD variants to affected children over siblings.With 298 opportunities, 31 more are transmitted only to the affected child rather than to the unaffected sibling among quad families (P = 0.0073 by permutation test).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingcitation

    All gene classes enriched for autism targets show decreased load of UR LGD variants, especially FMRP-associated.

    “All gene classes enriched for autism targets show a decreased load of UR LGD variants, especially the FMRP-associated class, as previously noted (10).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [10] source

      Not checked in the free run.

  • May combine two claims.

    Findingown data

    Differential transmission is 76 from mothers and 25 from fathers, with P = 0.0031 and P = 0.1890, respectively.

    “We observe differential transmission of 76 from the mother compared with 25 from the father with P = 0.0031 and P = 0.1890, respectively.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The two databases have similar total numbers of ultrarare synonymous variants and similar ratios of missense, synonymous, nonsense, and LGD mutations, as well as similar transition/transversion ratios.

    “The total number of UR synonymous variants is similar, as are the ratios of missense, synonymous, nonsense, and LGD mutations, as well as the ratios of transition and transversion UR variants.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Causal class vulnerability estimates for true autism genes based on recurrent, lower-IQ, and all affected classes are close (0.17, 0.19, and 0.15) and quite low.

    “The estimates of the causal class vulnerability for the true autism genes based on the rec in aut, autL, and aut values are close (0.17, 0.19 and 0.15, respectively) and quite low.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingreasoning

    The tolerance index is 1.0 for unaffected siblings and 0.15 for affected children.

    “The indices are 1.0 for unaffected siblings, by definition, and 0.15 for affected children.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingreasoning

    After adjusting for recurrently hit genes, 44% of the lower-IQ affected gene set and 30% of the higher-IQ gene set are estimated to be causal targets.

    “Adjusting for recurrently hit genes (Methods), we estimate that 44% of the gene set for lower IQ affected individuals and 30% for higher IQ individuals are causal targets.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Affected-child target classes have a markedly lower load of ultrarare LGD variants than sibling classes, though the difference is not significant for higher-IQ affected children.

    “Target classes from affected children show a markedly lower load for UR LGD variants than the class from siblings, although this difference is not significant for the affected children of higher IQ.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    For four of these genes transmissions mainly from mother; for KMT2C all four from father.

    “For four of these genes, the transmissions are mainly from the mother, but for one, the histone methyltransferase KMT2C, each of the four transmissions is from the father (Fig. S2).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Nearly all differential transmission signal comes from longer genes, with a differential of 93 in longer genes and 8 in shorter genes.

    “Despite this even split, nearly all the signal in differential transmission comes from the longer genes: the differential is 93 in the longer genes and 8 in the shorter genes.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Proportion of genes with reduced load matches ascertainment differential, with target genes in two classes.

    “Moreover, the proportion of genes with a reduced load matches expectations based on the ascertainment differential in DN mutation frequency between affected and unaffected siblings, with target genes falling into two classes: those genes with reduced mutational load and those genes without (Fig. 2).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Recently published RVIS findings show maternal transmission for lower half of RVIS-tolerant genes, and our result is similar (P=0.003 all, P>0.0005 long).

    “Recently published findings using the RVIS (16) present evidence for maternal transmission within the SSC of “private” LGD mutations for the lower half of RVIS-tolerant genes, and our result on maternal transmission based purely on genes with UR LGD mutations is very similar (statistically significant with a P = 0.003 for all, and P > 0.0005 for the more vulnerable long genes).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [16] source

      Not checked in the free run.

  • May combine two claims.

    Findingown data

    Proportion of causal genes in lower IQ is 44%, higher IQ 30%.

    “Similarly, we computed the proportion of causal genes in the targets of DN LGD mutations in affected children with lower (44%) and higher (30%) nonverbal IQs.”
    Quote in the paperNot a substantial claim
    • The authors state this as an assumption: Noncontributory DN LGD mutations in affected children occur at same rate as DN LGD mutations in unaffected individuals, and their target genes have similar UR LGD loads.
    • The authors state this as an assumption: Vulnerability of noncausal genes is 1, i.e., they have same load as targets of DN LGD mutation in unaffected siblings.

  • May combine two claims.

    Findingown data

    LOWESS fit agrees with square root model for UR LGD vs synonymous, linear for missense.

    “The LOWESS fit agrees closely with the square root model for the number of UR LGD mutations as a function of the number of UR synonymous variants, but it aligns better with the linear fit for the number of UR missense as a function of UR synonymous variants.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Study design and conductown data

    The affected and unaffected target classes were found on the same sequencing platforms, at the same depth, through identical pipelines, and in full siblings matched for germ-line background.

    “These two target classes were discovered on the same sequencing platforms, sequenced to the same depth, and processed through identical informatics pipelines, and they arise in children matched for germ-line background by being full siblings.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductown data

    Transmission significance was assessed by permuting affected labels.

    “We determined significance by permuting”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationcitation

    Because ASD individuals have reduced fecundity, target genes should be under severe purifying selection and have reduced load of disruptive variants.

    “Because individuals with ASD have drastically reduced fecundity (11), the net prediction is that these target genes should be under severe purifying selection, and hence have a reduced load of disruptive variants in the human gene pool.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [11] source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    There are 39 recurrent DN LGD targets in affected children, and about 90% are true causal targets.

    “There are 39 such targets (10,18), and we previously estimated that ∼90% of these targets are true causal targets.”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [10] source

      Not checked in the free run.

    • cites [18] source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation
    § Low load for disruptive mutations in autism genes and their biased transmission

    Genes that carry de novo likely gene-disruptive mutations in children with autism spectrum disorder are highly vulnerable, and a disruptive mutation in many vulnerable autism genes is likely to lead to autism spectrum disorder.

    “We previously computed that genes with de novo (DN) likely gene-disruptive (LGD) mutations in children with autism spectrum disorders (ASD) have high vulnerability: disruptive mutations in many of these genes, the vulnerable autism genes, will have a high likelihood of resulting in ASD.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [10] source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Recurrence and overlap analysis of de novo likely gene-disruptive targets estimates about 500 causative autism spectrum disorder target genes in lower-IQ affected individuals, enriched in certain functional classes.

    “From recurrence and overlap analysis of DN LGD targets, we estimate ∼500 causative ASD target genes in the affected individuals with lower intelligence quotients (IQ), and these targets are enriched in certain functional classes (10).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [10] source

      Not checked in the free run.

  • May combine two claims.

    Interpretationcitation
    § Low load for disruptive mutations in autism genes and their biased transmission

    Disruptive mutations in autism genes face strong negative selection and are rapidly removed from the population.

    “Because individuals with ASD have lower fecundity, such mutations in autism genes would be under strong negative selection pressure.”
    • Also stated in: § autism spectrum disorder | gene vulnerability | disruptive mutations | biased transmission | autism genes · boosted
      “Because people with ASD have lower fecundity than the general population, a disruptive mutation in an autism gene will be under strong purifying selection and quickly eliminated from the population (11).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [11] source

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Study design and conductnone given

    Some SSC parents may carry autism-contributing variation, so statistics from these families may overestimate rare variation in candidate autism genes.

    “Obviously, some of these parents may be carriers of variation contributing to ASD, so statistics extracted from these families may overestimate the rare variation seen in candidate autism genes.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    Some SSC parents may carry ultrarare LGD mutations ascertained because they caused autism in their affected children after transmission.

    “It is possible that some SSC parents carry UR LGD mutations that have been ascertained because they caused autism in their affected children upon transmission.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Interpretationreasoning

    Because most missense mutations are weakly selected or neutral, missense accumulation should be proportional to synonymous mutation.

    “Because the great majority of missense mutations will be under weak to neutral selection, we expect accumulation of missense will be proportional to synonymous mutation (Fig. 1B).”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductown data

    Significance of LGD ratio differences was assessed with 10,000 permutations that randomly swapped genes between sets while keeping set sizes fixed.

    “To determine if these lower ratios were significant, we performed 10,000 permutations, randomly swapping genes between sets while keeping the size of sets fixed, and computed the LGD ratio.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Half siblings sharing the maternal germ line have fourfold greater concordance than those sharing the paternal germ line.

    “for example, that there is a fourfold greater concordance in half siblings sharing the maternal rather than paternal germ-line (17).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [17] source

      Not checked in the free run.

  • May combine two claims.

    Study design and conductown data

    SSC exome data came from similar capture and sequencing platforms, a uniform analysis pipeline, and known coverage at every nucleotide position in targeted regions.

    “The advantage of using the SSC is that the WES data were obtained with similar capture and sequencing platforms and were subject to a uniform analysis pipeline, and the coverage is known for every nucleotide position in every gene in every person within the targeted regions.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Interpretationnone given

    Association of load with causation provides tool to study transmission of causative factors.

    “The association of load with causation provides us with a tool by which we can study transmission of causative factors.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationreasoning

    Power of using gene tolerance should be of general value in genetic disorders that reduce fecundity.

    “The power of using gene tolerance for disruptive mutation should be of general value in the analysis of genetic disorders that reduce fecundity.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationown data

    Our results strengthen this finding with stronger statistics.

    “Our results strengthen this finding with somewhat stronger statistics by using a larger set of target genes divided by severity, by using a larger control population to hone the tolerance score, or both.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Earlier population studies indicated importance of shared maternal bloodline in sibling risk.

    “Earlier population studies have indicated the importance of a shared maternal bloodline in sibling risk (19,20).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [19] source

      Not checked in the free run.

    • cites [20] source

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Premisereasoning

    We relied solely on LGD variants because their presence in interior of well-annotated gene will almost certainly disrupt function.

    “We have relied solely on LGD variants because, unlike missense variants that are hard to interpret, their presence in the interior of a well-annotated gene (Fig. S1) will almost certainly cause disruption of gene function.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationreasoning

    UR variants of vulnerable genes would be short-lived and not detectable by population association studies.

    “By this theory, UR variants of vulnerable genes would be short-lived in the population, and therefore not detectable by population association studies.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductnone given

    Score should not be treated as absolute probability; it is heuristic.

    “Such a score should not be treated as absolute probability for a gene to be an autism gene.It is a heuristic that allows us to prioritize genes and has several useful properties.”
    Quote in the paperNot a substantial claim
    • The authors state this as an assumption: Autism vulnerable genes and intolerant (protected) genes are treated as the same set.
    • The authors state this as an assumption: DN and rare parental data are independent.
    • The authors state this as an assumption: A protected gene has only 10% of typical share of UR LGD mutations.

  • Classified as probably not a substantial claim.

    Premisecitation

    Autism candidate genes have reduced load of damaging mutation, reported earlier.

    “That autism candidate genes have a reduced load of damaging mutation has been reported earlier by others using different methods for measuring tolerance of mutation based largely on missense mutation and/or overlapping autism sample sets (13,14,16).”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites [13] source

      Not checked in the free run.

    • cites [14] source

      Not checked in the free run.

    • cites [16] source

      Not checked in the free run.

37 of 37

  • The paper relies on this without support.

    The majority of LGD mutations, and the subset of nonsense mutations, are harmful.

    “The majority of LGD mutations (and the subset of nonsense) will be harmful, but this scenario is not so for missense mutations.”

  • The authors state this as an assumption.

    Noncontributory DN LGD mutations in affected children occur at same rate as DN LGD mutations in unaffected individuals, and their target genes have similar UR LGD loads.

    “The rate of noncontributory DN LGD mutations per affected individual should match the rate of DN LGD mutations in unaffected individuals; moreover, the genes targeted by noncontributory DN LGD mutations in affected and unaffected siblings should have similar loads of UR LGD mutations.”

  • The authors state this as an assumption.

    Vulnerability of noncausal genes is 1, i.e., they have same load as targets of DN LGD mutation in unaffected siblings.

    “where the vulnerability of noncausal genes is set to 1, reflecting the assumption that they have the same load as the targets of DN LGD mutation in unaffected siblings.”

  • The authors state this as an assumption.

    Autism vulnerable genes and intolerant (protected) genes are treated as the same set.

    “We developed a simple heuristic gene score based on the naive assumption that the set of autism vulnerable genes and the set of intolerant (protected) genes are the same.”

  • The authors state this as an assumption.

    DN and rare parental data are independent.

    “Using the Bayes rule, and further assuming that the DN and rare parental data are independent, we can rewrite this conditional probability as:”

  • The authors state this as an assumption.

    A protected gene has only 10% of typical share of UR LGD mutations.

    “To define pðRjAÞ, we assume that a protected gene has only 10% of the typical share of UR LGD mutations:”
6 of 6

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Exome sequencing of families has found evidence that inherited protein-truncating variants contribute to ASD.

    “Evidence for the contribution of inherited protein truncating variants to ASD has also been found by exome sequencing of families (Iossifov et al., 2015; Krumm et al., 2015)”

    From Getting to the cores of autism (2019) · cites this paper as Iossifov et al., 2015

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Higher-IQ transmission difference is reported as both 30 and 31

    Minor

    The prose gives a transmission difference of 30 for higher-IQ families, while Table 3 gives 31 for the same group.

  • Long-gene P value conflicts with the stated inequality

    Minor

    The discussion says the P value for the more vulnerable long genes is greater than 0.0005, but Table 3 reports it as equal to 0.0005.

  • Count of demoted recurrent targets changes from 11 to 12

    Minor

    The text says 11 targets were demoted, then refers to five of those targets as being among 12 genes.

  • Causal vulnerability for all affected targets has two values

    Minor

    The text reports 0.16 for causal targets across all affected individuals, whereas Table 2 reports 0.15 for that same class.

4 of 4

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Related

Authors
Iossifov, I., Levy, D., Allen, J., Ye, K., Ronemus, M., Lee, Y.-H., Yamrom, B., & Wigler, M.
Affiliations
  • Cold Spring Harbor Laboratory (US)
  • Albert Einstein College of Medicine (US)
  • Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 11724;
ORCID
Venue
Proceedings of the National Academy of Sciences
Publisher
National Academy of Sciences
Published
2015-09-23
Volume
112
Issue
41
Pages
E5600-7
ISSN
  • 0027-8424 (ISSN-L)
  • 1091-6490
DOI
10.1073/pnas.1516376112
PMC
PMC4611648
OpenAlex
W1961487924
Semantic Scholar
72f81b98f4b3861f39ce50b51ba3e5ca838b57a4
Cited by
156
References
21
Citations by year
2026: 11 · 2025: 7 · 2024: 13 · 2023: 11 · 2022: 11 · 2021: 18 · 2020: 23 · 2019: 18 · 2018: 16 · 2017: 14 · 2016: 11 · 2015: 2
FWCI
8.52
Type
journal article
Language
en
Open access
Licence
www.pnas.org/site/misc/userlicense.xhtml (vor)
Locations
  • Proceedings of the National Academy of Sciences (publishedVersion)
  • Europe PMC (PubMed Central) (publishedVersion)
  • PubMed Central (publishedVersion)
  • PubMed (publishedVersion)
  • Cold Spring Harbor Laboratory Institutional Repository (Cold Spring Harbor Laboratory) (acceptedVersion)
Funders
SFARI (SF235988, SF362665)
Topics
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Autism Spectrum Disorder Research
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Genetics › Genomics and Rare Diseases
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Genetics › Genetics and Neurodevelopmental Disorders
Keywords
autism spectrum disorder, de novo mutations, mutational load, mother-to-child transmission, fecundity, intelligence quotient, exome sequencing
MeSH terms
Autistic Disorder, Child, Child, Preschool, Female, Gene Pool, Humans, Male, Models, Genetic, Mutation, Databases, Genetic, Exome

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-06.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-06.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review