Everything is heritable: some new work on autism.

"The contribution of de novo coding mutations to autism spectrum disorder", Iossifov et al 2014 https://pdf.yt/d/fOlLOcil-4C_xKDJ / https://www.dropbox.com/s/690qberi1ecnne4/2014-iossifov.pdf / http://libgen.org/scimag/get.php?doi=10.1038%2Fnature13908

Whole exome sequencing has proven to be a powerful tool for understanding the genetic architecture of human disease. Here we apply it to more than 2,500 simplex families, each having a child with an autistic spectrum disorder. By comparing affected to unaffected siblings, we show that 13% of de novo missense mutations and 43% of de novo likely gene-disrupting (LGD) mutations contribute to 12% and 9% of diagnoses, respectively. Including copy number variants, coding de novo mutations contribute to about 30% of all simplex and 45% of female diagnoses. Almost all LGD mutations occur opposite wild-type alleles. LGD targets in affected females significantly overlap the targets in males of lower intelligence quotient (IQ), but neither overlaps significantly with targets in males of higher IQ. We estimate that LGD mutation in about 400 genes can contribute to the joint class of affected females and males of lower IQ, with an overlapping and similar number of genes vulnerable to contributory missense mutation. LGD targets in the joint class overlap with published targets for intellectual disability and schizophrenia, and are enriched for chromatin modifiers, FMRP-associated genes and embryonically expressed genes. Most of the significance for the latter comes from affected females.

Autism spectrum disorder (ASD) is characterized by impaired social interaction and communication, repetitive behaviour and restricted interests. It has a strong male bias, especially in high-functioning affected individuals. The contribution from transmission has long been suspected from increased sibling risk 1 , but more recently the role of germline de novo (DN) mutation has been established, first from large-scale copy number variation in simplex families 2-5 , and subsequently from exome sequencing. The smaller DN variants observed by DNA sequencing pinpoint candidate gene targets 6-8 . These developments have promoted a new model for causation, and re-evaluation of sibling risk 9,10 . Here we report whole exome sequencing of the Simons Simplex Collection (SSC) 11 and an extensive list of DN mutated targets, including 27 recurrent LGD (nonsense, frameshift and splice site) targets

We report on 2,517 of ,2,800 SSC families including ,800 that were previously published 6-8 . We sequenced 2,508 affected children, 1,911 unaffected siblings and the parents of each family. Within the SSC, the overall gender bias in affected individuals, 7 males to 1 female, is nearly twice that typically reported.

Previous studies presented evidence of functional clustering in targets of DN LGD mutation in affected individuals 6-8,16 . Our larger data set was examined with an improved null 'length model' for mutation in which the probability of DN mutation in a gene is proportional to its length (Methods and Extended Data Fig. 5). We tested for enrichment within DN LGD and missense targets in probands and siblings for the following six classes: (1) FMRP target genes, with transcripts bound by the fragile X mental retardation protein 8,17 ; (2) genes encoding chromatin modifiers; (3) genes expressed preferentially in embryos 18,19 ; (4) genes encoding postsynaptic density proteins 20 ; (5) essential genes 21 ; and (6) genes identified as Mendelian disease genes 22 (Table 1, Supplementary Table 6 and Methods). These data provide the strongest evidence yet for overlap of DN LGD targets in affected probands with FMRP targets (55 observed versus 34.1 expected; P= 4 3 10 24 ) and chromatin modifiers (26 observed versus 11.8 expected; P= 3 3 10 24 ). We also observed signal from mutation in genes expressed in embryonic development 23 (65 observed versus 45.0 expected; P= 2 3 10 23 ). The latter signal comes mainly from the small number of female affected individuals (23 observed versus 8.5 expected from 67 LGD targets; P= 5 3 10 26 ). The 27 genes with recurrent LGDs show strong enrichment for FMRP targets (14 observed versus 2.6 expected; P= 4 3 10 28 ) and chromatin modifiers (6 observed versus 0.9 expected; P= 2 3 10 24 ). By contrast, no significant enrichment for these gene sets is seen for the DN LGD targets in unaffected siblings. The 1,500 DN missense targets in probands are also enriched for FMRP targets and embryonically expressed genes. We observe 171 FMRP targets (144.8 expected; P= 0.03), and 220 embryonically expressed genes (191.4 expected; P= 0.03). As before, the signal for embryonically expressed genes comes almost entirely from the small number of female affected individuals (48 observed, 31.1 expected from 244 targets; P= 0.002). With the exception of chromatin modifiers, contributory DN missense and LGD mutations tend to strike similar functional classes of genes.
...Our analysis of functional clustering and overlaps within target classes suggests that the mutations ascertained in probands target restricted sets of vulnerable genes. We next sought evidence for excess recurrence of targets. We first examined synonymous mutations and mutations in unaffected children. Among the 647 synonymous events in probands, there are 25 gene targets found in more than one child, close to the null expectation of 19.9 (P = 0.13). Recurrent LGD (n 5 3 out of 179 events) or missense targets (70 out of 1,143 events) in unaffected siblings are also close to null expectations (P = 0.2 and 0.04, respectively). In affected males with higher IQ there are no excess recurrent targets among 137 LGDs mutations (2 observed, 1.0 expected, P = 0.3) or among 728 missense mutations (26 observed, 24.7 expected, P= 0.4). By contrast, among probands the number of recurrent LGD (n = 27 out of 391 events) and missense targets (145 out of 1,675 events) are not compatible with the null expectation of 7.6 (P < 0.0001) and 115.0 (P = 0.001), respectively. Given these findings, as well as the lack of overlap between targets of higher and lower IQ males, we focused on the joint class of female probands and affected males of lower IQ. For the joint class, there were 22 recurrent LGD targets among 254 events with 3.3 expected (P = 0.0001). For the 944 missense events, 60 recurrent targets are observed with 40.2 expected (P = 0.0005).

De novo mutation and IQ Higher IQ probands are heavily skewed towards males 24 . For further analyses, we chose to divide the affected male population roughly in half into higher and lower IQ sets. We investigated whether higher IQ (.90) males comprise a population with a distinguishable genetic signature. There is a decreased ascertainment differential for DN LGD mutations in male children with higher IQ relative to other affected individuals (Extended Data Fig. 3 and Supplementary Table 6). This is not statistically significant over the joint 403 region. However, over the entire data set, the drop in IQ is 5 points for males with DN LGD mutation compared to those without mutation (P= 0.01; Fig. 2). The mean IQ of affected males with recurrent DN LGDs drops 20 points (P= 0.00001, Fig. 2). Significance is also evident as we examine targets by functional class. Males with LGD mutations in FMRP targets have an average 14-point drop (P= 0.001). This trend continues with LGD targets in the other functional classes-chromatin modifiers and embryonically expressed genes-but with reduced significance. We observe little signal from DN missense mutation, even in recurrent targets, either because these events are less likely to contribute or because they are less severe. Female probands show the same trends as males, but as they comprise a smaller population, the significance is weak (Fig. 2).

The most likely number of genes vulnerable to DN mutations in the joint class is estimated to be 387 for LGD targets with a 95% confidence interval of 149- 915, and 404 for DN missense targets (confidence interval 71-3,050). From the length model and our estimate that 43% of LGD mutations are contributory, we have 90% confidence that a given LGD mutation contributes to autism in a gene recurrently hit by an LGD mutation (Methods).
...The SSC was assembled with the explicit hypothesis that finding targets of DN mutation would be a path to gene discovery. We now have 353 candidate LGD gene targets, 27 genes recurrently hit by LGD events, and 145 recurrent missense targets, each with about 40%, 90% and 35% chance of being contributory, respectively.

Targets in autism are enriched in certain functional categories, providing deeper support for previously published observations 6-8 . FMRPassociated genes and chromatin modifiers are prominent targets in all groups except higher IQ males. The former are thought to function in neuroplasticity. Embryonically expressed genes are significantly enriched as LGD or missense targets, but only in females. Enrichment in these genes may reflect that these contributory mutations cause alterations before a female protective effect takes place.
Recurrent LGD targets encode receptors, ion channels and synaptic proteins likely to function directly in neuro-circuitry (for example, SCN2A, GRIN2B and RIMS1), but also proteins functioning in cytoskeletal remodelling (for example, ANK2 and MED13L) and transcriptional regulation. Chromodomain helicase gene family members carry many recurrent LGDs. The most frequently hit gene is CHD8 (ref. 30), followed by CHD2 (three LGDs) and four other members (one LGD each) of that family. CHD8 is a transcriptional regulator thought to be important for suppression of the Wnt-b-catenin signalling pathway through histone H1 recruitment 31 . Another intriguing target is the protein kinase DYRK1A, hit four times and located in the Down's syndrome critical region 7 .

We can compute a distribution of class vulnerability for all vulnerable genes targeted by a given mutational type (Methods) because F, A and P have empirically sampled distributions and H has a distribution inferred from the total length of the gene class. The distribution of class vulnerability for DN LGDs in males with lower IQ has a mode around 0.4 (Fig. 3). In other words, ,40% of DN LGDs in vulnerable genes in a male contribute to diagnoses of lower IQ ASD. Similarly, ,10% of missense mutations in vulnerable genes contribute to diagnoses of lower IQ autism (Fig. 3). The mode for LGD vulnerability in females is fourfold lower than for lower IQ males, mainly because the prevalence is fourfold lower. Reduced penetrance in females is not well understood, but may be consequent to sexually dimorphic development. Support for this is seen in the relative enrichment of embryonically expressed genes as targets in females.

From the therapeutic perspective, the good news is that in almost all cases DN mutations occur in probands in whom a normal allele is also present. It is theoretically possible that enhancing activity of the remaining alleles might alleviate symptoms. So in our view, the long-term prognosis for treating ASD is positive. Moreover, ASD targets overlap with targets for intellectual disability and schizophrenia, so mechanism-based treatments might work for different diagnostic categories. In the intermediate term, functional clustering suggests that treatments might be tailored to a smaller number of convergent pathways.

DN substitutions increase ,0.4 per paternal decade (Extended Data Fig. 4), consistent with previous studies 15 and the increase in autism as a function of paternal age 46,47 . Where we could determine parental phase, DN substitutions arose more frequently in the paternal (287) than in the maternal (80) background. Among phased DN events, the mean age at birth was 34.6 for the father and 32.0 years for the mother, whereas the respective mean ages were 33.2 and 31.1 years for fathers and mothers in the whole population (P = 0.0001 and 0.047, respectively, that these differences arise by chance).
[see also "Extended Data Figure 4 | Paternal age and DN mutation rate at child birth." pg13: ~0.9 exomic substitutions at paternal age 25, rising linearly to ~1.6 at paternal age 48]

- "Synaptic, transcriptional and chromatin genes disrupted in autism", Rubeis et al 2014 https://pdf.yt/d/VEBB-iru3Nvzwo_D / https://www.dropbox.com/s/vg9pw9osi6rwewr/2014-rubeis.pdf / http://libgen.org/scimag/get.php?doi=10.1038%2Fnature13772

The genetic architecture of autism spectrum disorder involves the interplay of common and rare variants and their impact on hundreds of genes. Using exome sequencing, here we show that analysis of rare coding variation in 3,871 autism cases and 9,937 ancestry-matched or parental controls implicates 22 autosomal genes at a false discovery rate (FDR) , 0.05, plus a set of 107 autosomal genes strongly enriched for those likely to affect risk (FDR , 0.30). These 107 genes, which show unusual evolutionary constraint against mutations, incur de novo loss-of-function mutations in over 5% of autistic subjects. Many of the genes implicated encode proteins for synaptic formation, transcriptional regulation and chromatin-remodelling pathways. These include voltage-gated ion channels regulating the propagation of action potentials, pacemaking and excitability-transcription coupling, as well as histone-modifying enzymes and chromatin remodellers-most prominently those that mediate post-translational lysine methylation/demethylation modifications of histones.

Features of subjects with autism spectrum disorder (ASD) include compromised social communication and interaction. Because the bulk of risk arises from de novo and inherited genetic variation 1-10 , characterizing which genes are involved informs ASD neurobiology and reveals part of what makes us social beings.
Whole-exome sequencing (WES) studies have proved fruitful in uncovering risk-conferring variation, especially by enumerating de novo variation, which is sufficiently rare that recurrent mutations in a gene provide strong evidence for a causal link to ASD. De novo loss-of-function (LoF) single-nucleotide variants (SNVs) or insertion/deletion (indel) variants 11-15 are found in 6.7% more ASD subjects than in matched controls and implicate nine genes from the first 1,000 ASD subjects analysed 11-16 . Moreover, because there are hundreds of genes involved in ASD risk, ongoing WES studies should identify additional ASD genes as an almost linear function of increasing sample size 11 .
Here we conduct the largest ASD WES study so far, analysing 16 sample sets comprising 15,480 DNA samples (Supplementary Table 1 and Extended Data Fig. 1). Unlike earlier WES studies, we do not rely solely on counting de novo LoF variants, rather we use novel statistical methods to assess association for autosomal genes by integrating de novo, inherited and case-control LoF counts, as well as de novo missense variants predicted to be damaging.

Some 13.8% of the 2,270 ASD trios (two parents and one affected child) carried a de novo LoF mutation-significantly in excess of both the expected value 19 (8.6%, P , 10 214 ) and what was observed in 510 control trios (7.1%, P = 1.6 3 10 25 ) collected here and previously published 15 . Eighteen genes (Table 1) exhibited two or more de novo LoF mutations...we would expect to observe this in approximately two such genes by chance. While we expect only two de novo Mis3 events in these 18 genes, we observe 16 (P = 9.2 3 10 211 , Poisson test). Because most of our data exist in cases and controls and because we observed an additional excess of transmitted LoF events in the 18 genes, it is evident that the optimal analytical framework must involve an integration of de novo mutation with variants observed in cases and controls and transmitted or untransmitted from carrier parents.

TADA identified 33 autosomal genes with an FDR , 0.1 (Table 1) and 107 with an FDR , 0.3 (Supplementary Tables 2 and 3 and Extended Data Fig. 3). Of the 33 genes, 15 (45.5%) are known ASD risk genes 9 ; 11 have been reported previously with mutations in ASD patients but were not classed as true risk genes owing to insufficient evidence (SUV420H1 (refs 11, 15), ADNP 12 , BCL11A 15 , CACNA2D3 (refs 15, 21), CTTNBP2 (ref. 15), GABRB3 (ref. 21), CDC42BPB 13 , APH1A 14 , NR3C2 (ref. 15), SETD5 (refs 14, 22) and TRIO 11 ) and 7 are completely novel (ASH1L, MLL3 (also known as KMT2C), ETFB, NAA15, MYO9B, MIB1 and VIL1). ADNP mutations have recently been identified in 10 patients with ASD and other shared clinical features 23 . Two of the newly discovered genes, ASH1L and MLL3, converge on chromatin remodelling. MYO9B plays a key role in dendritic arborization 24 . MIB1 encodes an E3 ubiquitin ligase critical for neurogenesis 25 and is regulated by miR-137 (ref. 26), a microRNA that regulates neuronal maturation and is implicated in schizophrenia risk 27 .
When the WES data from genes with an FDR , 0.3 were evaluated for the presence of deletion copy number variants (CNVs) (such CNVs are functionally equivalent to LoF mutations), 34 CNVs meeting quality and frequency constraints (Supplementary Information) were detected in 5,781 samples (Extended Data Fig. 1). Of the 33 genes with an FDR , 0.1, 3 contained deletion CNVs mapping to 3 ASD subjects and one parent. Of the 74 genes meeting the criterion 0.1 # FDR , 0.3, about one-third could be false positives. Deletion CNVs were found in 14 of these genes and the data supported risk status for 10 of them (Extended Data Table 1 and Extended Data Fig. 4). Two of these ten, NRXN1 and SHANK3, were previously implicated in ASD 2,3,10 . The risk from deletion CNVs, as measured by the odds ratio, is comparable to that from LoF SNVs in cases versus controls or transmission of LoF variants from parents to offspring.

By analysing the distribution of relative risk over inferred ASD genes 20 , the number of ASD risk genes can be estimated. The estimate relies on the balance of genes with multiple de novo LoF mutations versus those with only one: the larger the number of ASD genes, the greater proportion that will show only one de novo LoF. This approach yields an estimate of 1,150 ASD genes (Supplementary Information).