The causes of autism spectrum disorder or ASD are not fully understood; researchers believe both genetics and environment play a role.
In some cases, the disorder is linked to de novo mutations that appear only in the child and are not inherited from either parent’s DNA.
In a study published December 23, 2019 in Nature Medicine, an international team of scientists, led by researchers at University of California San Diego School of Medicine, describe a method to measure disease-causing mutations found only in the sperm of the father, providing a more accurate assessment of ASD risk in future children.
“Autism afflicts 1 in 59 children and we know that a significant portion is caused by these de novo DNA mutations, yet we are still blind to when and where these mutations will occur,” said co-senior author Jonathan Sebat, PhD, professor and chief of the Beyster Center for Molecular Genomics of Neuropsychiatric Diseases at UC San Diego School of Medicine.
“With our new study, we can trace some of these mutations back to the father, and we can directly assess the risk of these same mutations occurring again in future children.”
Recent studies suggest gene-damaging de novo mutations are involved in at least 10 to 30 percent of ASD cases, with the number of mutations rising with the father’s age at time of conception.
De novo mutations occur spontaneously in parents’ sperm or eggs or during fertilization. The mutation is then present in each cell as the fertilized egg divides.
Studies now point to male sperm as a particularly important source of these mutations, with the chance of the mutation recurring within the same family generally estimated at 1 to 3 percent.
“However, such estimates are not based on actual knowledge of the risk in an individual family, but instead are based on frequencies in the general population,” said co-senior study author Joseph Gleeson, MD, Rady Professor of Neuroscience at UC San Diego School of Medicine and director of neuroscience research at the Rady Children’s Institute for Genomic Medicine. “When a disease-causing mutation occurs for the first time in a family, the probability that it could happen again in future offspring is not known. Thus families must make a decision with a great deal of uncertainty.”
For their study, Gleeson, Sebat and colleagues analyzed the sperm of eight fathers who were already parents of children with ASD.
The goal was to look for the presence of multiple, genetically different material in cells in the same person, a phenomenon called mosaicism.
Using deep whole genome sequencing, they found variants in offspring that were matched only in the fathers’ sperm.
“While medical textbooks teach us that every cell in the body has an identical copy of DNA, this is fundamentally not correct.
Mutations occur every time a cell divides, so no two cells in the body are genetically identical,” said first author Martin Breuss, PhD, an assistant project scientist in Gleeson’s lab.
“Mosaicism can cause cancer or can be silent in the body. If a mutation occurs early in development, then it will be shared by many cells within the body.
But if a mutation happens just in sperm, then it can show up in a future child but not cause any disease in the father.”
In this illustration of sperm mosaicism, mutated sperm are depicted in red. Image is credited to UCSD.
The researchers determined that disease-causing mutations were present in up to 15 percent of the fathers’ sperm cells, information that could not be determined through other means, such as blood samples.
“My laboratory has a long-standing interest in understanding the origins of pediatric brain disease, and how mutations contributes to disease in a child,” said Gleeson.
“We previously showed that mosaicism in a child can lead to diseases like epilepsy. Here, we show that mosaicism in one of parents is at least as important when thinking about genetic counseling.”
If developed into a clinical test, the researchers said fathers could have their sperm studied to determine their precise risk of recurrence in future children. The methods might also be applied to men that haven’t had children yet, but who want to know the risk of having a child with a disease.
Disclosure: Martin Breuss, Danny Antaki, Morgan Kleiber, Kiely N. James, William M. Brandler, Jonathan Sebat and Joseph Gleeson are inventors on a provisional patent (PCT Ref. No. SD2017-181-2PCT) filed by UC San Diego and titled “Assessing risk of de novo mutations in males.”
Co-authors include: Danny Antaki, Morgan Kleiber, Oanh Hong, Madhusudan Gujral, William M. Brandler, Ileena Mitra, and Melissa Gymrek, UC San Diego; Renee D. George, Kiely N. James, Laurel L. Ball, Xiaoxu Yang, Sara A. Wirth, Jing Gu, Camila A. B. Garcia, Damir Musaev, An Nguyen, Jennifer McEvoy-Venneri, and Evan Sticca, Howard Hughes Medical Institute at UC San Diego and Rady Children’s Institute for Genomics of Psychiatric Diseases; Renatta Knox, Howard Hughes Medical Institute at UC San Diego, Rady Children’s Institute for Genomics of Psychiatric Diseases and Weill Cornell Medical College; Martha Cristina Cancino Botello and Javiera Uribe Fenner, Hospital Universitari Mutua de Terrassa, Spain; Maria Cárcel Pérez, Maria Arranz, and Amaia Hervás, Fundacio Docencia i Recerca Mutua Terrassa, Spain; Andrea B. Moffitt and Zihua Wang, Cold Spring Harbor; and Orrin Devinsky, New York University School of Medicine.
And that would be the best holiday gift of all.
Autism Spectrum Disorder (ASD) is a complex biological condition characterized by a common set of behavioral features with early childhood onset, reflecting the interaction between different genetic and environmental risk factors [1].
At present, there is no ultimate treatment for the core features of ASD. Nevertheless, autistic symptoms can be reduced by early behavioral interventions [2,3], and some pharmacological therapies are available for the treatment of psychiatric comorbidities [4].
ASD prevalence seems to be increasing: most recent estimates suggest a prevalence of 1 in 59 among 8-year-old children from the USA (https://www.cdc.gov/ncbddd/autism/data.html) [5]. Another study estimated a 3.5 prevalence increase between 2001 and 2011 in 2- to 17-year-old children [6].
What caused this increased prevalence, beyond a broadening of ASD diagnostic criteria and a better ascertainment of cases, is still unclear. Still, as ASD is the final consequence of cascade events impacting brain development from gestation to early postnatal life [7], it is possible that a true rise is related to these complex events.
While the etiology of ASD is not fully understood, genetics is a well-established risk factor [8]. Twin studies suggested a 76% concordance in monozygotic twins, confirming a strong genetic hereditability for ASD, but also supporting an important contribution of environmental factors [9].
Genetic defects in more than 100 genes and loci, and hundreds of copy number variants (CNVs) and single nucleotide (SNVs) polymorphisms (SNPs) have been implicated in about 20% of ASD cases [10,11,12,13]. DNA microarrays enable the discovery of rare and recurrent CNVs as important contributors to ASD and lead to gains in the understanding of autism genetics and to the identification of individuals who might be genetically susceptible to autism. Hotspots of recurrent CNVs, including 16p11.2, 22q11.2, 1q21.1, 7q11.23, and 15q11–q13, have been shown to be strongly associated with ASD [14].
Next-generation sequencing (NGS) methods revolutionized ASD gene discovery and have also substantially contributed to functional genetic data, linking mutations frequently associated with ASD with genes involved in the regulation of brain transcriptional networks during brain development and early synaptogenesis, thus throwing some light on the understanding of the neurobiological consequences of the disruption of these ASD-associated genes [12,15]. Nevertheless, also single-genes syndromes have been associated with ASD, including Fragile-X (FMR1), Tuberous Sclerosis Complex (TSC1-2) and PTEN syndrome [16,17].
Nonetheless, the heterogeneous clinical and biological phenotypes observed in ASD strongly suggest that, in genetically susceptible individuals, environmental risk factors also combine or synergize to generate a “threshold point” that might determine a dysfunction. While progress has been made towards gaining an understanding of genetic and epigenetic factors, environmental risk factors are less understood [18].
Actually, recent studies have demonstrated that during critical periods of central nervous system development, early exposure to a variety of environmental factors, ranging from microbes (bacteria and viruses) to medications, from chemicals to physical agents, can affect neurobiological development, including effects relevant to ASD [19,20].
In October 2018, international ASD experts convened in Rome to discuss the potential pathogenetic role of environmental factors, as well as their interactions with genetic susceptibility, focusing on three biologically relevant windows for brain development: the periconception, prenatal and early postnatal periods.
From the epidemiological point of view, the identification of the exact timing of action of each environmental factor, as well as its consequences in the neurodevelopmental pathways, remains elusive. Nevertheless, it is now possible to establish some differentiations among risk factors that can assist in developing detection and personalized follow-up of populations at higher risk for ASD.
In this paper, we summarize the results of this consensus conference and put forward clinical recommendations for clinicians to reduce ASD risk and/or its severity.Go to:
Conception Period
Advanced parental age: The association between older parental age at conception and neuropsychiatric disorders in offspring is now well documented [21,22]. In the case of ASD, both advanced maternal and paternal age at time of birth (≥35 years) were associated with an increased risk of ASD [23,24,25,26].
Emerging evidence also confirms a combined parental age effect, which is highest when both parents are in the older age range and increases with increasing differences in parental ages [27]. Both human and animal model studies support the hypothesis of an association between elevated rates of de novo mutations in older fathers and increased risk of ASD [28,29].
It has been also suggested that maternal mechanisms mediating the effects of advanced maternal age on ASD risk are associated not only with chromosomal or genetic modifications, but also with a higher prevalence of chronic diseases and a less favorable uterine environment, often resulting in more obstetrical complications, which might eventually lead to an increased risk of adverse birth outcomes [26].
Use of hormonal induction and/or assisted reproductive technologies: Assisted Reproductive Technologies (ART) now account for 1–3% of all live births in the Western world (https://www.cdc.gov/reproductivehealth/index.html) [30]. Several procedures that are used in the ART process, such as hormonal stimulation, egg retrieval, in vitro fertilization (IVF), intra-cytoplasmic sperm injection (ICSI), micro-manipulation of gametes and exposure to culture medium, could subject the gametes and early embryos to environmental stress and may be associated with an increased risk of birth defects and low birth weight (LBW) [31].
Children conceived using ART are also at a higher risk for congenital anomalies including a two-fold increase in the central nervous system and epigenetic and imprinted disorders [32,33,34]; there is some evidence that ART might have an impact on imprinting through DNA methylation [35]. Actually, assisted conception and ASD share several risk factors. In both cases, hormonal disturbances, especially in testosterone/androgen regulation, along with high rates of advanced parental ages, preterm deliveries, and LBW, have been reported [6,24,36,37].
Additionally, a recent meta-analysis indicated that the use of ART may be associated with a higher risk of ASD in the offspring [31]. In a previous case-control study conducted on a large Israeli population [38], a higher ART prevalence (IVF and ICSI) (10.7%) even in young mothers (<29 years) was reported among ASD children compared to the overall ART rate.
In addition, the study ruled out the hypothesis that ART was associated with unique autism symptomatology (i.e. autism severity and adaptive functioning, a history of developmental regression) that may represent a distinct clinical phenotype in this group.
The study results indicated that although assisted conception may be a risk factor for ASD, this group did not appear to represent a separate clinical phenotype within the autism spectrum. These findings suggest that the increased recent prevalence of both ART and ASD might be related.
Environmental chemical and toxicant factors: There is some evidence that exposure to chemical pollutants at critical developmental stages may affect neural and behavioral development. The pathogenetic mechanisms of environmental chemical factors can involve neurotoxicity but can also extend to pathways of immune dysregulation, altered lipid metabolism, and mitochondrial dysfunction. To date, the strongest evidence of association is shown by traffic-related air pollutants and pesticides at different times of exposures [39,40].
Maternal nutritional status: Maternal nutritional status and body mass index before pregnancy have been considered as environmental factors that can influence normal brain development through excess or deficit of micronutrients and growth factors, which can affect neurodevelopmental outcomes of offspring [41,42]. In this view, both maternal obesity and underweight have been associated with an increased risk of ASD [42,43].
Maternal obesity results in activation of the maternal immune system and in a chronic inflammation of the uterine environment potentiating abnormal neuronal growth and differentiation in the fetus, with consequent neurodevelopmental impairments in the offspring [44]. At the same level, maternal undernutrition may elicit a physiological stress response leading to neuronal damage through a disproportionate release of proinflammatory factors [45].
A large number of recent studies have suggested association between pre-conception intake of folate and risk of ASD onset in newborns [46]. A significantly higher rate of ASD has been found in children not exposed to folic acid (FA) compared to in children of mothers who took it.
Conversely, some apparently conflicting results were reported by other studies that related an increased risk for ASD and neurocognitive impairments in children of mothers who used dietary supplements of synthetic FA [47,48,49]. A possible explanation of these diverging results might be offered by the different compositions between the FA used in supplements (pteroylmonoglutamic) and the one from natural food sources (ormyl-tetrahydropteroylglutamates).
High levels of pteroylmonoglutamic acid, which depend on liver-based metabolism, could result in high levels of unmetabolized and non-useful FA in the blood, which can cause changes in brain synaptic transmission and dysregulation of expression of many genes associated with ASD [50,51,52].
Another important micronutrient potentially linked to the neurodevelopmental alterations in ASD is iron. The importance of a correct intake of iron is evident already from the peri-conception period [53]. In the brain, iron contributes to neurotransmitter production, myelination and immune function. In this view, iron deficiency in this period could result not only in impairment in the general development of cognitive, motor and language skills, but also in deficit in social orientation and engagement that could lead to ASD [53].
Medications: A growing number of researches highlighted the potential association of prenatal exposure to Selective Serotonin Reuptake Inhibitors (SSRIs) with the onset of ASD, hypothesizing a pathogenetic link between alterations in serotonin pathways and ASD neurobiological abnormalities [54,55,56]; exposure during the preconception period or the first trimester seems to be associated with a higher risk compared to the other two trimesters [57]. Others have found that antidepressants, regardless of their composition, might be associated with increased ASD risk [58]. Thus, some diverging results have been found in relation to both antidepressant types and dosages [58,59].
Furthermore, a Danish longitudinal study, with a follow-up of 5,057,282 person-years, did not detect a significant association between maternal use of SSRIs during pregnancy and ASD in the offspring [60]. Moreover, another large research did not find, after controlling for several confounding factors, a significant association between prenatal exposure to antidepressant medication and ASD [61].
Lastly, a “confounding by indication” cannot be excluded, raising the possibility that it is depression and anxiety that might be risk factors for offspring ASD, rather than antidepressants per se [58]. Nevertheless, even if evidence is still conflicting, the recommendation is to proceed to apply the precautionary principle, balancing the use of antidepressants against the substantial adverse consequences of untreated maternal depression.Go to:
Prenatal Period
Environmental chemicals and toxicants: In the last few years, epidemiologic investigations indicated that prenatal exposure to chemical and toxic factors such as air pollution, pesticides, materials used in the plastic industry and heavy metals may increase the risk of ASD [39,40,62,63,64]. Possible mechanisms behind the association between these environmental risk factors and ASD are not only their interactions with genetic factors, and/or epigenetic marks leading to a diminished ability to detoxify xenobiotics [65,66] but also their potential role in triggering neuro-inflammation and oxidative stress that lead to neurobiological and neurotransmitter alterations and abnormalities in signaling pathways [63].
Air pollution: Air pollution is probably the chemical risk factor with the strongest evidence of association with ASD, especially for exposures in the third trimester [40,67,68]. Multiple variables, such as metrics of exposure, type of pollution, time of exposure, could influence the risk of ASD and its clinical outcome [69,70,71].
It should be noted, however, that despite positive associations that were observed in many countries like the USA, Canada, Taiwan and Israel, European studies did not find any association [71,72,73]. Recently, in a Canadian population-based birth cohort, a significant association between exposure to nitric oxide and ASD was found. No association was found between ASD and particulate matter with a diameter of <2.5 μm or nitrogen dioxide [74].
These contradictory results might be due to the fact that studies of air pollution have been often limited by indirect and cross-sectional methods of exposure measurement, by different metrics of exposure, by different evaluations of outcomes, and by focusing on different pollutants. Notwithstanding the need for further investigation, and even if some unanswered questions remain, prenatal air pollution exposure has emerged as a potentially modifiable risk factor for ASD.
Pesticides: Exposure to organochlorine pesticides (measured using geographical mapping) increases the risk of offspring ASD [75,76]. Additionally, studies that examined risk of ASD in relation to prenatal levels of poly-chlorinated biphenyls (PCBs) reported a suggestive association with specific PCBs [20,77,78]. Organophosphate exposure during pregnancy increases the risk of autistic symptoms in the offspring, at 2–3 years of age [79]. In particular, residential proximity to organophosphates at some point during pregnancy is associated with a 60% increased risk for offspring to develop ASD [80].
Conversely, a pilot case-control study investigating risk associated with exposure to organic pollutants (including a variety of PCB congeners, DDT – dichloro-diphenyl-trichloroethane, and DDE – Dichloro-2,2-bis(p-chlorophenyl)-ethylene) measured in archived maternal serum and diagnosis of ASD in children did not find significant differences in odds ratios for ASD [77].
Most pesticides of current use are neurotoxic, may target the developing brain [81] and are prone to cause oxidative stress [82]. The widely used pyrethroids have been associated to ASD and neurodevelopmental delay [83]. Nevertheless, for the same reasons described before, when talking about air pollution, an association between pesticide exposure and ASD is not yet confirmed [78].
Phthalates: Phthalates are a class of chemicals used as plasticizers, solvents, and lubricants, and as enteric coatings on pharmaceuticals and nutritional supplements. Few studies have addressed the relationship between ASD and prenatal exposure to phthalates (3rd trimester) with contrasting results [78].
Heavy metal exposure: Little evidence for an association between hair metal concentration of mercury, copper, cadmium, selenium, chromium and autistic symptoms has emerged until now [84]. Moreover, as most of these studies only measured biomarkers and do not ascertain actual exposure sources, temporality of association is unknown.
Some studies examined exposure in relation to maternal dental amalgam fillings and maternal or child consumption of seafood with inconsistent findings [19]. A meta-analysis found not only consistent evidence for lack of association between childhood thimerosal exposure and ASD, but also an increased risk of ASD following a higher level of inorganic mercury exposure [85].
Medications: The association between ASD and prenatal exposure to drugs is increasingly investigated; a specific area of interest was the study of antiepileptic and antidepressant agents [54].
Among antiepileptic drugs (AED), valproate showed the strongest association with neurodevelopmental outcome, in terms of cognitive disabilities, developmental delay, and ASD [86]. It is therefore contraindicated as a first-line antiepileptic or mood stabilizer in pregnant women or in those who plan pregnancy. Moreover, other AED, as oxcarbazepine and lamotrigine (alone or combined with valproate), have been found to be associated with the onset of ASD in the offspring [87].
Findings across several meta-analyses examining the association between antidepressant exposure during pregnancy and ASD are reasonably consistent showing an increased risk [88]. Additionally, maternal psychiatric disorders could play a critical role in the development of ASD; thus, these have been considered also as a potential confounding or addictive risk factor for exposure to antidepressants alone [88,89,90].
Some studies suggested also a possible link between prenatal or early-life antibiotic use and ASD [91], but too limited information is currently available to draw conclusions. Recently, however, it has been demonstrated that low-dose antibiotic exposure in late pregnancy and early postnatal life in mice induces impaired social behaviors and aggression in mice associated with changes in the intestinal microbiome [92].
On the other hand, supplementation with the probiotic Lactobacillus Rhamnosus JB-1 might prevent the early-life antibiotic-induced aberrant behaviors. Taken together, these results merit further research on the potential role of early-life antibiotic exposure in the development of ASD.
Substance abuse: A large number of studies examined prenatal exposure to substance abuse as heavy tobacco smoke, alcohol, or cocaine and ASD. Association between high amounts of alcohol consumption in pregnancy and ASD in offspring (especially those with Fetal Alcohol Syndrome) is documented [93,94,95]. On the other hand, association between moderate alcohol intake in pregnancy and ASD is unlikely [94].
An association between smoking during pregnancy and risk of childhood autism has been suggested [96], but in this case, results are conflicting, with two meta-analyses in a total of 15 studies reporting no association with overlapping odds ratios [97,98]. Therefore, at present, insufficient data have been found to support an association.
Nutritional factors: Epidemiological studies and data obtained in humans have provided evidence that mother’s diet during pregnancy plays a critical role in the development of the neural circuitry that regulates behavior, thus determining persistent behavioral effects in the offspring [48]. Generally, it is known that some elements of maternal diet during pregnancy, such as FA, vitamin D, iron and fatty acids, are associated with higher or lower incidence of ASD or autistic traits in the offspring [99].
Specifically, low concentrations of vitamin D and FA are associated with an increased risk of ASD diagnosis, in particular if these deficiencies are present in the mid-gestational period [100,101]. In addition, a maternal diet with high levels of methanol and aspartame during gestation could be linked to an increased risk of ASD [102].
A poor omega-3 intake during gestation and maternal high-fat diet during pregnancy has been associated with the risk of ASD and other neurodevelopmental disorders [19,103]. In fact, high-fat consumption during pregnancy is strongly associated with activation of several of the same inflammatory cytokines (e.g., interleukins IL-4, and IL-5) that are elevated during gestation in mothers of children with ASD.
Furthermore, high-fat diet consumption in pregnant women is associated with modifications of the neural pathways involved in behavioral regulation, specifically the serotoninergic system. The suppression of serotoninergic synthesis in the brain may underlie the risk of developing later behavioral disorders, as long as the offspring is exposed to maternal high-calorie diet during pregnancy.
Prenatal infections and maternal immune activation: Current data suggest that at least for a subset of women, exposure to infections during pregnancy might increase ASD risk or other disorders of the central nervous system (CNS) in the offspring. Activation of the maternal immune response can confer a risk for the onset of psychiatric disorders. In particular, exposure to prenatal infections, such as flu, rubella, measles, herpes simplex virus, and bacterial infections, may increase the risk for the offspring of developing bipolar disorder and schizophrenia [104].
More recently, some population-based cohort studies described a potential link between autism risk and maternal infection or inflammation during pregnancy, depending on the time of gestational exposure, the type of infective agent, and the intensity of the maternal immune response; specifically, viral infections seem to be associated to ASD risk in the first trimester, bacterial infections in the second trimester, influenza and febrile episodes during the whole pregnancy but especially in the third trimester [105,106]. Fewer studies have examined the potential impact on ASD risk of fever as such, rather than in connection with infection broadly [106].
A retrospective case-control study based on maternal self-report showed an association between fever during pregnancy and increased ASD risk [105]; it showed also that this risk was attenuated only in mothers who took anti-pyretic medications to control their fever, but not in those mothers who did not [105]. A prospective study in Norway also found an increased risk for ASD after prenatal fever exposure, as well as evidence of a dose–response relationship, with risks rising parallel to multiple episodes of maternal fever [107].
A prevailing concept is that maternal immune activation (MIA) may alter the expression of inflammatory molecules in the developing fetus and that maternal-fetal immune dysregulation may disrupt brain development and neural connectivity, which in turn may have long-term effects on the offspring’s mental functions [108]. Among the studies supporting a link between maternal infection and increased risk of ASD, there are several ones carried out with the quantification of cytokine, chemokines and of other inflammatory mediators measured in the maternal serum and amniotic fluid [109]. These studies, however, have generated conflicting results [56,105]. Recently, increased levels of maternal cytokines and chemokines during gestation have been associated with subsequent ASD with intellectual disability [110].
Maternal immune systems can be involved in increasing ASD risk, even independently from prenatal infections. In particular, maternal autoantibodies might recognize proteins in the developing fetal brain [111]; these autoantibodies can be detected in ~20% of mothers of children at risk for developing autism versus 1% of mothers of typically developing children, and defined an additional sub-phenotype of ASD [112,113].
Individual maternal factors and diseases: Gestational diabetes has been considered a risk factor because it negatively affects fetal growth and it increases the rate of pregnancy complications [114,115,116]. Moreover, it impacts long-term fine and gross motor development and leads to learning difficulties and attention-deficit hyperactivity disorder [117]. These adverse effects of maternal diabetes on brain development may arise from the increased fetal oxidative stress, as well as from epigenetic changes in the expression of several genes [114,115,118]. However, the increased risk for ASD linked to gestational diabetes may be related to pregnancy complications rather than to complications secondary to hyperglycemia. Whether control of diabetes reduces ASD risk is still unknown [114,115].
Additionally, maternal melatonin levels have been investigated as potential culprits in the ASD pathogenesis [119]. Melatonin is a crucial hormone for neurodevelopment and protects from oxidative stress and neurotoxicant agents. Melatonin deficiency is frequently detected in ASD children already in a very early period of life, and thus the potential implications of low maternal melatonin levels have been considered as a factor that might increase the susceptibility to autism [120].
Source:
UCSD
Media Contacts:
Scott LaFee – UCSD
Original Research: Closed access
“Autism risk in offspring can be assessed through quantification of male sperm mosaicisms”. Martin W. Breuss, Danny Antaki, Renee D. George, Morgan Kleiber, Kiely N. James, Laurel L. Ball, Oanh Hong, Ileena Mitra, Xiaoxu Yang, Sara A. Wirth, Jing Gu, Camila A. B. Garcia, Madhusudan Gujral, William M. Brandler, Damir Musaev, An Nguyen, Jennifer McEvoy-Venneri, Renatta Knox, Evan Sticca, Martha Cristina Cancino Botello, Javiera Uribe Fenner, Maria Cárcel Pérez, Maria Arranz, Andrea B. Moffitt, Zihua Wang, Amaia Hervás, Orrin Devinsky, Melissa Gymrek, Jonathan Sebat & Joseph G. Gleeso al.
Nature Medicine doi:10.1038/s41591-019-0711-03.
