+1 Recommend
0 collections
      • Record: found
      • Abstract: found
      • Article: not found

      A method and server for predicting damaging missense mutations


      Read this article at

          There is no author summary for this article yet. Authors can add summaries to their articles on ScienceOpen to make them more accessible to a non-specialist audience.


          To the Editor: Applications of rapidly advancing sequencing technologies exacerbate the need to interpret individual sequence variants. Sequencing of phenotyped clinical subjects will soon become a method of choice in studies of the genetic causes of Mendelian and complex diseases. New exon capture techniques will direct sequencing efforts towards the most informative and easily interpretable protein-coding fraction of the genome. Thus, the demand for computational predictions of the impact of protein sequence variants will continue to grow. Here we present a new method and the corresponding software tool, PolyPhen-2 (http://genetics.bwh.harvard.edu/pph2/), which is different from the early tool PolyPhen1 in the set of predictive features, alignment pipeline, and the method of classification (Fig. 1a). PolyPhen-2 uses eight sequence-based and three structure-based predictive features (Supplementary Table 1) which were selected automatically by an iterative greedy algorithm (Supplementary Methods). Majority of these features involve comparison of a property of the wild-type (ancestral, normal) allele and the corresponding property of the mutant (derived, disease-causing) allele, which together define an amino acid replacement. Most informative features characterize how well the two human alleles fit into the pattern of amino acid replacements within the multiple sequence alignment of homologous proteins, how distant the protein harboring the first deviation from the human wild-type allele is from the human protein, and whether the mutant allele originated at a hypermutable site2. The alignment pipeline selects the set of homologous sequences for the analysis using a clustering algorithm and then constructs and refines their multiple alignment (Supplementary Fig. 1). The functional significance of an allele replacement is predicted from its individual features (Supplementary Figs. 2–4) by Naïve Bayes classifier (Supplementary Methods). We used two pairs of datasets to train and test PolyPhen-2. We compiled the first pair, HumDiv, from all 3,155 damaging alleles with known effects on the molecular function causing human Mendelian diseases, present in the UniProt database, together with 6,321 differences between human proteins and their closely related mammalian homologs, assumed to be non-damaging (Supplementary Methods). The second pair, HumVar3, consists of all the 13,032 human disease-causing mutations from UniProt, together with 8,946 human nsSNPs without annotated involvement in disease, which were treated as non-damaging. We found that PolyPhen-2 performance, as presented by its receiver operating characteristic curves, was consistently superior compared to PolyPhen (Fig. 1b) and it also compared favorably with the three other popular prediction tools4–6 (Fig. 1c). For a false positive rate of 20%, PolyPhen-2 achieves the rate of true positive predictions of 92% and 73% on HumDiv and HumVar, respectively (Supplementary Table 2). One reason for a lower accuracy of predictions on HumVar is that nsSNPs assumed to be non-damaging in HumVar contain a sizable fraction of mildly deleterious alleles. In contrast, most of amino acid replacements assumed non-damaging in HumDiv must be close to selective neutrality. Because alleles that are even mildly but unconditionally deleterious cannot be fixed in the evolving lineage, no method based on comparative sequence analysis is ideal for discriminating between drastically and mildly deleterious mutations, which are assigned to the opposite categories in HumVar. Another reason is that HumDiv uses an extra criterion to avoid possible erroneous annotations of damaging mutations. For a mutation, PolyPhen-2 calculates Naïve Bayes posterior probability that this mutation is damaging and reports estimates of false positive (the chance that the mutation is classified as damaging when it is in fact non-damaging) and true positive (the chance that the mutation is classified as damaging when it is indeed damaging) rates. A mutation is also appraised qualitatively, as benign, possibly damaging, or probably damaging (Supplementary Methods). The user can choose between HumDiv- and HumVar-trained PolyPhen-2. Diagnostics of Mendelian diseases requires distinguishing mutations with drastic effects from all the remaining human variation, including abundant mildly deleterious alleles. Thus, HumVar-trained PolyPhen-2 should be used for this task. In contrast, HumDiv-trained PolyPhen-2 should be used for evaluating rare alleles at loci potentially involved in complex phenotypes, dense mapping of regions identified by genome-wide association studies, and analysis of natural selection from sequence data, where even mildly deleterious alleles must be treated as damaging. Supplementary Material 1

          Related collections

          Most cited references6

          • Record: found
          • Abstract: found
          • Article: not found

          Human non-synonymous SNPs: server and survey.

          Human single nucleotide polymorphisms (SNPs) represent the most frequent type of human population DNA variation. One of the main goals of SNP research is to understand the genetics of the human phenotype variation and especially the genetic basis of human complex diseases. Non-synonymous coding SNPs (nsSNPs) comprise a group of SNPs that, together with SNPs in regulatory regions, are believed to have the highest impact on phenotype. Here we present a World Wide Web server to predict the effect of an nsSNP on protein structure and function. The prediction method enabled analysis of the publicly available SNP database HGVbase, which gave rise to a dataset of nsSNPs with predicted functionality. The dataset was further used to compare the effect of various structural and functional characteristics of amino acid substitutions responsible for phenotypic display of nsSNPs. We also studied the dependence of selective pressure on the structural and functional properties of proteins. We found that in our dataset the selection pressure against deleterious SNPs depends on the molecular function of the protein, although it is insensitive to several other protein features considered. The strongest selective pressure was detected for proteins involved in transcription regulation.
            • Record: found
            • Abstract: found
            • Article: found
            Is Open Access

            SNPs3D: Candidate gene and SNP selection for association studies

            Background The relationship between disease susceptibility and genetic variation is complex, and many different types of data are relevant. We describe a web resource and database that provides and integrates as much information as possible on disease/gene relationships at the molecular level. Description The resource has three primary modules. One module identifies which genes are candidates for involvement in a specified disease. A second module provides information about the relationships between sets of candidate genes. The third module analyzes the likely impact of non-synonymous SNPs on protein function. Disease/candidate gene relationships and gene-gene relationships are derived from the literature using simple but effective text profiling. SNP/protein function relationships are derived by two methods, one using principles of protein structure and stability, the other based on sequence conservation. Entries for each gene include a number of links to other data, such as expression profiles, pathway context, mouse knockout information and papers. Gene-gene interactions are presented in an interactive graphical interface, providing rapid access to the underlying information, as well as convenient navigation through the network. Use of the resource is illustrated with aspects of the inflammatory response and hypertension. Conclusion The combination of SNP impact analysis, a knowledge based network of gene relationships and candidate genes, and access to a wide range of data and literature allow a user to quickly assimilate available information, and so develop models of gene-pathway-disease interaction.
              • Record: found
              • Abstract: found
              • Article: found
              Is Open Access

              Increased Susceptibility to Cortical Spreading Depression in the Mouse Model of Familial Hemiplegic Migraine Type 2

              Introduction Migraine is a clinically heterogeneous disorder affecting more than 10% of the general population. It generally occurs with unilateral and pulsating severe headache often accompanied by nausea, photophobia and phonophobia. In approximately one third of migraineurs, the headache attack is preceded by aura, a transient neurological symptom that are most frequently visual but may involve other senses [1]. The migraine attack is triggered by a brain dysfunction that leads to activation and sensitization of the trigeminovascular system, particularly trigeminal nociceptive afferents innervating the meninges and lastly to headache [2], [3], [4]. Neuroimaging examination suggests that migraine aura is associated to cortical spreading depression (CSD), a short-lasting, intense wave of neuronal and glial cell depolarization. CSD spreads slowly over the cortex at a rate of approximately 2–5 mm/min and is followed by long lasting depression of neuronal activity [5], [6], [7], [8]. Experimental evidence on patients and animal models supports CSD as both underlying migraine aura [1], [7], [8], [9] and a key triggering event for trigeminal activation [10], [11], [12], although the role of CSD in migraine headache is still debated. As an indirect confirmation, several migraine prophylactic agents cause an increase of CSD initiation threshold [13]. Common migraine has a strong multifactorial genetic component, which is higher in migraine with aura (MA) than in migraine without aura (MO) [14], [15]. As for many other multifactorial diseases whose complexity hampers the investigation of the pathogenetic mechanisms, rare monogenic forms that phenocopy most or all the clinical features of the common disease are of great help for describing the complicated events leading to migraine. Familial hemiplegic migraine (FHM) is a rare autosomal dominant subtype of MA, whose aura symptoms include hemiparesis. Aura symptoms and headache duration are usually longer in FHM than MA, but all other headache properties are similar. FHM is genetically heterogeneous and is associated to mutations in three different genes. Mutations in CACNA1A [16], ATP1A2 [17] and SCN1A [18] genes are responsible for Familial hemiplegic migraine type 1 (FHM1), type 2 (FHM2), and type 3 (FHM3), respectively. The CACNA1A and SCN1A genes both encode neuronal voltage-gated ion channels, whereas the ATP1A2 gene encodes the α2 subunit of the Na,K-ATPase, hence suggesting a key role of cation trafficking in the pathophysiology of FHM. Until now, more than 50 FHM2 mutations have been identified and most of these are missense mutations. A small fraction of mutations is represented by microdeletions [19] and a single mutation affecting the stop codon, which causes an extension of the ATP1A2 protein by 27 aminoacid residues [20]. Most of the ATP1A2 mutations are associated with pure FHM without additional clinical symptoms [17], [19], [20], [21], [22]. However, a number of FHM2 mutations have been associated to complications like cerebellar ataxia [23], childhood convulsions [24], epilepsy [25] and mental retardation [26]. Interestingly, ATP1A2 mutations associated with non-hemiplegic migraine phenotypes, such as basilar migraine and even common migraine have been reported [27], [28]. The Na,K ATPase is a P-type ion pump that utilizes the free energy of ATP hydrolysis to exchange Na+ for K+ and maintains gross cellular homeostasis. The functional pump is a heterodimer, consisting of one α catalytic subunit and one β subunit that is required for protein folding, assembling, membrane-addressing, and modulates substrate affinity [29]. The α subunit exposes both the amino- and carboxy- termini in the cytoplasm and crosses the plasma membrane with ten transmembrane segments (M1–M10) [30]. Four isoforms of α Na,K-ATPase (α1, α2, α3 and α4) are present in mammals [29], [31]. While no pathogenic mutations are known for the ubiquitous α1- and the testis α4-subunits, mutation in both α2 and α3 isoforms cause neurological diseases when mutated, FHM2 and rapid-onset dystonia parkinsonism, respectively [32]. While in the adult brain the α1 isoform is nonspecifically present in both neurons and glial cells and α3 is neuron-specific, the α2 isoform is essentially expressed in astrocytes [33]. Investigation of the functional consequences of FHM2 mutations in heterologous expression systems revealed that these mutations produce partial or complete loss of function of the α2 Na,K pump [34], [35], [36]. Here, we report the generation of the first mouse model of FHM type 2, a knock-in mutant harboring the W887R ATP1A2 mutation. The W887R mutation localizes to the extracellular loop between M7 and M8, which includes the β subunit binding site [37] and was shown to produce the almost complete loss of pump activity [17], [38]. Homozygous Atp1a2R887/R887 mutants die just after birth, while heterozygous Atp1a2+/R887 mice are fertile and show no apparent clinical phenotype. However, heterozygous FHM2 mouse displays altered CSD properties, such as decreased threshold and increased velocity of propagation. We hypothesize that inefficient astrocyte-mediated clearance of glutamate from the synaptic cleft is a key event for the enhanced susceptibility to CSD in the FHM2 mouse. Results Generation of FHM2 knock-in mutant mouse With the aim of investigating the molecular pathogenesis of FHM type 2, we generated a knock-in mouse model by inserting an FHM2 mutation, the transition T2763C that causes the aminoacid replacement W887R in the Atp1a2 murine gene (construct details in M&M). The amino acid sequence conservation between human and mouse α2 Na,K-ATPase proteins is very high and, in particular, in the extracellular domain between transmembrane domains M7–M8, where W887R is located [17]. This mutation was one of the first two mutations reported to be associated to typical cases of the disease. Embryonic stem cells harboring the R887 and the neo cassette were injected in C57Bl/6J blastocysts and then transferred to pseudopregnant CD1 females. We obtained three chimeric mice, one of which transmitted the Atp1a2+/R887-neo allele through germline (Figure 1A). Heterozygous Atp1a2+/R887-neo mice were genotyped by Southern blot analysis (Figure 1C), are fertile and display no apparent phenotype. To remove the neo cassette that hampers the natural expression of the mutant allele, we crossed the Atp1a2+/R887-neo mice with transgenic mice expressing the Flippase recombination enzyme (FLPe) under the control of the human ACTB promoter (TgN(ACTFLPe)9205Dym; The Jackson Laboratory). Hence, we obtained the heterozygous Atp1a2+/R887 knock-in mice (Figure 1B), which are fertile as well and show no visible clinical phenotype. Contrary to heterozygous mice, homozygous Atp1a2R887/R887 mutants do not survive beyond the first day post partum, thus resembling the neonatal lethal phenotype of the Atp1a2 null mutant [39], which succumbs for dysfunctional neuronal activity and respiratory distress. Therefore, we addressed our investigation onto the heterozygous knock-in mouse, which shares the Atp1a2 gene asset with FHM2 patient. The general behavior of heterozygous Atp1a2 +/R887 mice was tested by a modified SHIRPA protocol [40] that provides comparable quantitative data on animal motor, sensory, autonomic and neuropsychiatric functions. The scored parameters are summarized in Table 1. No major differences in the sensory-motor functions were observed between heterozygous Atp1a2+/R887 (n = 8) and wild-type (n = 6) mice, except for a higher fear and anxiety of Atp1a2+/R887 at the specific tests of transfer arousal and fear (p 90% congenic. Heterozygous Atp1a2 +/R88 and Atp1a2 +/+ littermates were used for further analysis. Behavioral analysis Sensory-motor function of mutant mice compared with controls was assessed by a modified version of the SHIRPA protocol primary screening [40]. Briefly, undisturbed behavior of each animal was first observed in its own home cage: body position, spontaneous activity and respiration rate were recorded, assigning a score to each behavior. In addition, manifestations of tremors, bizarre behaviors, stereotypes or convulsions were checked at this stage of the protocol. Thereafter mice were transferred individually to a new arena and were tested for transfer arousal, palpebral closing, piloerection, gait, pelvic and tail elevation, touch escape and positional passivity. There followed a sequence of manipulations using tail suspension and a grid across the width of the arena; animals were scored for trunk curl, limb grasping and grip strength. To complete the assessment, the animals were restrained in a supine position to record autonomic behaviors (heart rate, skin color, limb and abdominal tone, lacrimation, salivation) prior to measurement of the righting reflex after flip of the animal. Vocalizations and irritability (during supine restrain) were also recorded. Fear was assessed based on reaction to transfer to a new environment. A score was assigned to each behavioral test as described in Table 1. RT-PCR Total RNA was extracted from embryonic mice (E19.5) (n = 9, 3 embryos for each genotype) neuronal (brain) tissues by Trizol method (Invitrogen, Carlsbad, CA, USA). RNA was reverse transcribed using random hexamers SuperScript® First-Strand Synthesis System (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. Atp1a2 cDNA was amplified using forward primer on exon 19 (5′-GGCTTCTTTACCTACTTTGTGATA-3′) and reverse primer on exon 20 (5′-ATGCCCTGCTGGAACACTGAGTTG-3′) with Hot Master Taq DNA polymerase (Eppendorf, Hamburg, Germany) at 94°C for 2 min, 35 cycles at 94°C for 30 s, 58°C for 30 s, 65°C for 30 s, and 65°C for 5 min. This strategy allows amplification of both endogenous wild-type and mutant allele (PCR product: 254 bp). The relative Atp1a2 amount was normalized to the β-actin expression levels (610 bp PCR product). Since the R887 missense mutation introduces a new restriction site for MspI enzyme, the PCR product was subsequently digested with MspI (New England Biolabs, Ipswich, MA, USA) to discriminate the endogenous gene (uncut, band size: 254 bp band) and the mutant (cut, bands size: 178 bp+76 bp). PCR products were run on a 2% agarose gel in TAE buffer. Western blot analysis To prevent proteolysis during the procedure, all steps were carried out on ice, and all buffers contained protease inhibitor cocktail (Roche, Mannheim, Germany) and phenylmethanesulfonyl fluoride (1 mM). Embryonic brains of the various genotypes (n = 12, 4 for each genotype) were processed simultaneously. For the extraction of membrane proteins, whole brain was homogenized with a glass-Teflon homogenizer in Sucrose solution (0.32 M Sucrose, 5 mM Hepes pH 7.4, 2 mM EDTA). After a short centrifugation (5000 rpm, 20′4°C) the supernatant was centrifuged for 1 hr at 42,000 rpm 1 h 4°C (Beckman, ultraTL100, rotor TL100.3) and the pellet resuspended in Sucrose buffer. Protein concentration was measured using the Bio-Rad Protein Assay according to the manufacturer's instructions. The preparation of cells and tissues (total brain, cortex and cerebellum) total lysates were performed adding RIPA buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.5% sodium deoxycholate, 0.1% SDS, 2 mM EDTA, and 1% Igepal CA-630) to the collected samples and left 30′ on ice than the lysates were centrifuged 13000 g 20′4°C. The protein content of the supernatant was measured using the BCA protein assay with bovine serum albumin as standard. We resuspended equal amounts of proteins (15 µg each sample in 20 µl) in SDS-PAGE buffer (100 mM Tris-Glycine pH 6.8, 0.56 M mercaptoethanol, 2% SDS, 15% glycerol, and 0.1% BFB), and separated them for 2 h at 100volt in 8% SDS-polyacrilamide gels. Proteins were electrophoretically transferred to hybond ECL nitrocellulose membranes (GE Healthcare, Munich, Germany) and blots were blocked overnight with 5% non-fat milk 0.1% Tween-20 in PBS. The blocked blots were incubated for 2 h with subunit specific antibodies, washed three times for 10 minutes each with 0.1% Tween-20 in PBS then incubated with the appropriate peroxidase conjugated secondary antibodies. After another series of washes (three times for 10 minutes each) peroxidase was detected using a chemiluminescent substrate (GE Healthcare, Munich, Germany). Transfections and immunocytochemistry Plasmid constructs were the same as in [17]) by metafectene (Biontex, Martinsried/Planegg, Germany) according to the manufacturer's instructions. We selected the ratio range of Metafectene (µl) to plasmids DNA (µg) of 5∶1. 48 h after transfection, we fixed HeLa cells in 4% paraformaldehyde (PFA) for 30 min at RT and blocked and permeabilized with 10% donkey serum 0.2% Triton-X100 in phosphate-buffered saline solution (PBS) for 30 min at RT. Permeabilized cells were then incubated with primary antibodies for 2 hr at RT, than washed (three times) in PBS, incubated with appropriate secondary antibodies and washed three times with PBS solution. We placed cells in fluorescent mounting medium (Dako Cytomation, Glostrup, Denmark) over microscope slides and confocal microscopy was performed on the Perkin Elmer UltraVIEW. Colocalization analysis Immunofluorescence colocalization was visualized by confocal microscopy and analyzed by Wright Cell Imaging Facility (WCIF) colocalization plug-in of Image J software (http://www.uhnresearch.ca/facilities/wcif/imagej/colour_analysis.htm). The following parameters were measured: Pearson's correlation coefficient (Rr; 1, perfect correlation, to −1, perfect exclusion); Mander's overlap coefficient (R; 1, highest, to 0, random correlation); Ch1∶Ch2, the red∶green pixel ratio. Proteasome inhibitor Proteasome inhibitor MG132 (carbobenzoxy-L-leucyl-L-leucyl-L-leucinal) was obtained from Sigma-Aldrich, Milan, Italy (cat. C2211). MG132 were dissolved in DMSO and applied to cells at the concentration of 10 µM, after 48 hours of transfection for the time periods indicated in the text and. An equivalent volume of DMSO was added to control cells. Anti ubiquitin antibody was used to reveal the increase of ubiquitinated proteins after proteasome inhibition. Cortical spreading depression CSD was recorded as described in Van den Maagdenberg, et al. [46]. Briefly, mice (20–30 g) were anaesthetized with urethane (20% in saline; 6 ml/kg i.p.). Animals, mounted on a stereotaxic apparatus were continuously monitored for adequate level of anesthesia, temperature, heart rate and nociceptive reflexes. Blood oxygen saturation and flux as well as heart and breathing rates were monitored non-invasively using an oximeter (Starr, Life Science Corp.). Oxygen was supplied to maintain blood oxygenation above 93% for the entire duration of the experiment. Heart rate was between 400–600 beats/min, and breathing rate approximately 200 breaths/min. Animals not meeting these criteria were excluded from our sample. To record CSD three holes were drilled in the skull over the left hemisphere. The first corresponded to the occipital cortex and was used for access of the electrical stimulation electrode (0 mm A-P, 2 mm M-L from lambda). The second hole, at the parietal cortex (1 mm M-L, 1 mm caudal to bregma) and the third hole, at the frontal cortex (1 mm M-L, 1 mm rostral to bregma), were used for placement of the CSD recording electrodes. The steady (DC) potential was recorded with glass micropipettes filled with NaCl (3 M, tip resistance 1–2 MΩ) inserted 200 µm below the dural surface. An Ag/AgCl reference electrode was placed subcutaneous above the nose. Cortical stimulation was conducted using a copper bipolar electrode (0.2 mm tip diameter, 0.3 mm intertip distance) placed on the cortex surface after removing the dura. Single pulses of increasing intensity (20, 30, 40, 50, 60, 80, 100, 120, 140, 160, 180, 200, 230, 260, 290, 320, 350, 380, 430, 480, 530, 600, 700, 800, 900, 1000 µA) were applied for 100 ms at 3-min intervals by using a stimulus isolator/constant current unit (WPI, USA) until a CSD event was observed [13]. DC cortical potential was amplified (10×) and low-pass filtered at 200 Hz (Cyberamp, Axon Instruments, Union City, CA). Signals were continuously digitized and recorded using Labview data acquisition and analysis system. The minimal stimulus intensity at which a CSD event was elicited was taken as the CSD threshold. In all mice, when CSD was elicited, recordings were continued for 90 min to detect multiple CSDs. To estimate CSD propagation velocity, the distance between the two recording electrodes was divided by the time elapsed between the CSD onsets at the first and second recording sites. The percentage of mice with multiple CSD events was determined only from the mice that could be recorded for one full 90 min following the first detected event. CSD duration was measured at half-maximal amplitude [13]. Because no difference in CSD threshold and propagation rate was observed between male (N = 11 wild type and N = 11 mutants) and female (N = 7 wild type and N = 9 mutants) within each genotype (wild type: threshold male 20.7±2.1 µC, female 18.7±3.5 Mann-Whitney test p = 0.61; propagation rate male 3.9±0.42 mm/min, female 3.8±0.66 mm/min Mann-Whitney test p = 0.86; mutants: threshold male 13.4±3.0 µC, female 12.6±1.3 t-test p = 0.82; propagation rate male 5.2±0.32 mm/min, female 5.6±0.84 mm/min Mann-Whitney test p = 0.94) the results from males and females were pooled. Statistical analysis For SHIRPA protocol primary screening, comparisons were performed with the Mann-Whitney nonparametric test. The Student's t-test with one-tail distribution was used for significance calculation in densitometric analysis. Statistical analysis for CSD recordings was performed using Sigma Stat 3.1 (Systat Software, Chicago IL USA). Multiple groups were compared by ANOVA followed by post-hoc comparisons applying Bonferroni correction or Holm-Sidak test. When two groups were compared, t-test was applied. Normality and homoschedasticity of the data was checked. Data not normally distributed were compared using nonparametric Kruskal-Wallis ANOVA or Mann-Whitney rank sum test. Significance level was equal to 0.05. Data are reported as average ± SEM.

                Author and article information

                Nat Methods
                Nature methods
                24 March 2010
                April 2010
                1 October 2010
                : 7
                : 4
                : 248-249
                [1 ]Division of Genetics, Brigham & Women’s Hospital, Harvard Medical School, Boston, Massachusetts, USA
                [2 ]Department of Biochemistry, Max Planck Institute for Developmental Biology, Tübingen, Germany
                [3 ]Department of Systems Biology, Harvard Medical School, Boston, Massachusetts, USA
                [4 ]Engelhardt Institute of Molecular Biology, Russian Academy of Sciences, Moscow, Russia
                [5 ]Life Sciences Institute and Department of Ecology and Evolutionary Biology, University of Michigan, Ann Arbor, USA
                [6 ]European Molecular Biology Laboratory, Heidelberg, Germany
                Author notes
                Correspondence to: Shamil R. Sunyaev 1 ssunyaev@ 123456rics.bwh.harvard.edu

                These authors contributed equally to this work


                Users may view, print, copy, download and text and data- mine the content in such documents, for the purposes of academic research, subject always to the full Conditions of use: http://www.nature.com/authors/editorial_policies/license.html#terms

                Funded by: National Institute of Mental Health : NIMH
                Funded by: National Institute of General Medical Sciences : NIGMS
                Award ID: R01 MH084676-02 ||MH
                Funded by: National Institute of Mental Health : NIMH
                Funded by: National Institute of General Medical Sciences : NIGMS
                Award ID: R01 GM078598-03 ||GM

                Life sciences
                Life sciences


                Comment on this article