Genetics of Autism

Many people are of the belief that there must be a genetic component to autism, and that the condition has "spontaneously" arisen due to some random gene mutation(s). Reasons cited are that they believe the condition is highly heritable in that if one identical twin has the condition, there is about an 80% chance that the other twin will have the condition. Similarly, there is a higher chance of autism in a family if there already is a child with autism. The problem with these theories is that they cannot explain the single generational increase in the rate of autism from one baby every thousand, to one baby in every 64 children. Such an increase in frequency can, though, be explained by an environmental change, and hence the possibility exists that the increase in the rate of autism may occur in individuals with suitable (or more correctly unsuitable or susceptible) genetics.

If there is a genetic component to the development of autism, then the possibility exists that selection occurred at the time of fertilization, and so the fertilized embryo was always destined to have autism, and as such would always have autism, and it would never resolve during a person's life-time. Such selection should be fairly uniform and hence one would expect to find a specific gene that caused the condition. Alternatively, there could be a nature vs nurture event in that particular gene variants may make the child more likely to develop the condition if it was raised in an unfavourable environment. In this case one would say that the child has a genetic predisposition to the condition, which is only "expressed" under certain environmental factors, such as poor nutrition in the womb. Another possibility is that certain genotypes may actually have a selective advantage in the womb, and it is only those that have the genotype that survive. What complicates the genetics is that if one poses that autism is due to more than one genetic factor, then one could have a combination of all the above scenarios, thus making it extremely hard to identify genetic susceptibility for autism.

Most genetic mutations occur at a single position (nucleotide) on the DNA that codes for a particular gene, and thereby cause a single nucleotide polymorphism (SNP) to be present. Some mutations do not affect the function of the protein that is coded for by the gene and others can have a profound affect on the protein, and can cause such a disruption in the protein structure that the protein is completely non-functional. If this occurs in a protein that is essential in the cell, it may result in death of the cell, and as such be conditionally lethal. In this case, the frequency of the double copy of the mutation will be zero. Additionally any one gene can have multiple mutations at different positions in the gene. In this case it is somewhat harder to determine what the effect will be on the protein. There are cases when some double mutations, whilst individually having a slight effect, may in combination become conditionally lethal.

An example of an environmentally "sensitive" gene is the gene coding for the Methylene tetrahydrofolate reductase gene, and the C677T variant selection. In low folate concentrations the less active 677T variant is strongly selected against and the frequency of the "TT" variant is greatly reduced when compared to that that occurs in a folate supplemented environment. If this 677TT variant is combined with the 1298CC and/or 1572GG variant this also appears to be conditionally lethal, as we could find no incidence of the dual combination occurring. For such environmentally sensitive genes, the incidence of the TT, CC or GG SNPs changes between societies supplementing with folate and those that don't. This observation suggests a selection against the lower activity variants of certain SNPs if the embryo is "raised" in a deficient environment.

One of the biggest puzzles in examining the genetics of autism is the observation that the incidence of autism is roughly four times as high in boys as it is in girls. Genetically the most logical place to look for a "risk" gene would therefore be on the X-Chromosome, as boys only have one copy of the chromosome, whereas girls have two copies. Under a single gene mutation event, in a non-selected environment, a girl would have to have two copies of the mutation (one on each of her X chromosomes) whilst the boy would only need one copy (as he only has one X chromosome). This would also imply that the mutated gene would have to come from the X chromosome of the mother. In order for a daughter to have ASD, both the mother and the father would need to have the mutation, whereas for the son, only the mother would need the mutation in order for it to be passed onto the son. The problem with this theory is that girls with autism would only be produced from parents in which the father had autism and the mother was a carrier. There is, however, no evidence that this occurs.

There is though the possibility that the father of a daughter with autism would carry the gene for susceptibility to the condition, but would not get the condition if the environment in the womb of the father's mother was did not elicit the condition. An example would be an insufficiency in vitamins or iron, particularly if these are causative for the condition.

Potentially relevant genes would be those that are susceptible to different levels of activity in different vitamin levels, iron levels or vitamin D levels, and our studies have shown that indeed there is a difference in SNP expression in ASD kids when compared to the normal population. These genes are discussed below.

Variation in vitamin D processing related SNPs in ASD

Potentially, if the development of ASD is associated with low vitamin D levels in the womb of the mother, then there should be a higher prevalence of genetic polymorphisms that one would associate with lower vitamin D. Recent advancements in DNA testing enables on to look at variants (polymorphisms) in single nucleotides that code for the proteins. These single nucleotide polymorphisms have been "coined" SNPs. We have examined the DNA from 80 individuals with ASD and compared their SNPs to those of 45 normal individuals of good health. Vitamin D processing occurs via a number of cytoplasmic enzymes called cytochromes, so called because they are coloured due to the presence of an iron atom bound within the protein in a heme-ring structure. A specific sub-group these proteins absorbs light at 450 nm, and as such have been named  the CyP450 proteins, or CYPs.

Cyp2R1 Vitamin D-25-hydroxylase

The vitamin D 25-hyroxylase enzyme, found in the liver, converts vitamin D3 to 25-hydroxyvitamin D (calcidiol). This form of the vitamin is inactive, but it is generally the form that is measured in serum.

The data represents the Percentage Risk +/+ compared between groups

SNP	Risk Allele	ASD +/+	ASD +/-	ASD -/-	Normal +/+	Normal +/-	Normal -/-
rs2060793	G=>.A	10.4	42.9	46.8	13.3	55.6	31.1
rs7041	A=>C	17.9	53.8	28.2	13.3	40.0	46.7
rs3829251	G=>A	9.0	35.9	55.1	0	29.8	70.2
rs7935792	A=>C	1.4	20.0	78.6	0	2.2	97.8

For each of the SNPs rs7041, rs3829251, and rs7935792, there was an increase in the relevant risk alleles, and a decrease in the protective alleles.

Cyp27B1 Vitamin D alpha-1-hydroxylase

The vitamin D alpha-1-hyroxylase enzyme, found in the kidney, converts 25-hydroxyvitamin D (calcidiol) to 1,25-dihydroxyvitamin D (calcitriol), which is the active form of the enzyme. Percentage Risk +/+ compared between groups

SNP	Risk Allele	ASD +/+	ASD +/-	ASD -/-	Normal +/+	Normal +/-	Normal -/-
rs4646536	A=>G	16.4	38.4	45.2	2.2	28.3	69.6
rs2296241	G=>A	26.0	53.2	20.8	23.9	50.0	26.1
rs703842	A=>G	16.7	38.5	44.9	2.2	26.1	71.7
rs2228570	G=>A	12.9	40.0	47.1	8.1	40.5	51.4
rs4516035	C=>.T	12.8	34.6	52.6	21.7	52.2	26.1
rs3087243	G=>A	19.2	57.7	23.1	23.9	45.7	30.4
rs2248359	C=>T	18.2	45.5	36.4	13.0	58.7	28.3
rs10876994	A=>C	10.0	35.7	54.3	2.4	16.7	81.0

For each of the SNPs rs4646536 (7.56 x), rs703842 (7.67 x), and rs10876994 (4.2 fold), there was an increase in the relevant risk alleles, and a decrease in the protective alleles. Interestingly both SNPs rs4646536 (7.56 x), rs703842 (7.67 x) have also been found to have a higher association with MS frequency, supporting the notion that low vitamin D is involved in ASD pathology (2).

Cyp24A1 Vitamin D 24-hydroxylase

The vitamin D 24-hyroxylase enzyme, found mainly in the kidney and placenta, is involved in the degradation of 1,25-dihydroxyvitamin D (calcitriol). More active variants of the enzyme would be expected to result in lower circulating 1,25-DiOHD. Percentage Risk +/+ compared between groups

SNP	Risk Allele	ASD +/+	ASD +/-	ASD -/-	Normal +/+	Normal +/-	Normal -/-
rs7975232	A=>C	20.5	50.0	29.5	21.7	50.0	28.3
rs12794714	G=>A	21.9	53.4	24.7	13.0	69.6	17.4
rs927650	C=>T	21.1	38.2	40.8	25.0	40.9	34.1
rs2248359	C=>T	17.1	44.7	38.2	15.9	56.8	27.3

There was little difference between ASD and control groups as far as frequency of any of the 4 SNPs studied.

Variation in neurotransmitter processing related SNPs in ASD

Potentially any behavioural differences seen in ASD could be due to genetic differences in neurotransmitter related genes, particularly those that are single step modifiers of amino acids, such as phenylalanine hydroxylase (PAH)(which converts phenylalanine to tyrosine) tyrosine hydroxylase (TH) which converts tyrosine to DOPA), tryptophan hydroxylase (TPH1 and TPH2)(converting tryptophan to serotonin), GAD (which synthesizes GABA from glutamine), and the serotonin transporter (SLC6A)

Phenylalanine-hydroxylase

This enzyme converts the essential amino acid phenylalanine to tyrosine, the DOPA precursore

SNP	Risk Allele	ASD +/+	ASD +/-	ASD -/-	Normal +/+	Normal +/-	Normal -/-
rs1522305	G=>C	1.4	31.9	66.9	4.4	35.6	60.0
rs3817446	T=>C	8.3	38.9	52.8	7.0	18.6	74.4
rs1718301	G=>A	18.1	43.1	38.9	20.9	48.8	30.2
rs11111419	A=>T	11.1	37.5	51.4	11.4	52.3	36.4

Overall there was a selection away from the minor SNPs, however there was a small increase in frequency of the rs3817446C allele in ASD.

Tyrosine-hydroxylase

This enzyme converts the amino acid tyrosine, the DOPA

SNP	Risk Allele	ASD +/+	ASD +/-	ASD -/-	Normal +/+	Normal +/-	Normal -/-
rs2070762	G=>A	29.2	50.0	20.8	19.6	41.3	39.1
rs28934581	T=>G	0	0	100	0	0	100
rs28934580	C=>T	0	0	100	0	0	100
rs6356	C=>T	23.6	47.2	29.2	13.3	46.7	40.0
rs28940881	A=>G	0	1.3	98.7	0	0	100

Three SNPs rs 28934581, rs28934580 and rs28940881 were found to have very, very low variation in SNP frequency suggesting that either these SNPs were conditionally lethal, or there has been a relatively recent appearance of the SNPs in the general population. There was a small shift towards the recessive alleles in both rs2070762 and rs6356 in ASD individuals potentially suggesting an association between the minor alleles and ASD. Many children though would have been regarded as "normal" for both alleles being "+/-", so the influence on the condition would be minor.

Tryptophan-hydroxylase

This enzyme converts the amino acid tryptophan to 5-hydroxytryptophan, the serotonin precursor. There are two forms TPH1, which works in the peripherary and bowel, and TPH2, which works in the brain.

SNP	Risk Allele	ASD +/+	ASD +/-	ASD -/-	Normal +/+	Normal +/-	Normal -/-
rs7130929 (TH1)	A=>C	15.5	49.3	35.2	6.4	53.2	40.4
rs4570625 (TH2)	G=>T	10.4	36.4	53.2	0	50	50
rs10506645	C=>T	4.2	40.3	55.6	2.1	44.7	53.2
rs4448731	T=>C	16.7	51.4	31.9	17.4	54.3	17.4
rs17110690	G=>A	2.8	38.9	58.3	4.3	43.5	52.2
rs11178997	T=>A	1.4	23.6	75.0	0	21.7	78.3
rs4760820	C=>G	12.5	50	37.5	10.9	37.0	52.2
rs1487275	A=>C	9.2	44.7	46.1	8.7	45.7	45.7
rs4565946	C=>T	16.9	48.1	35.1	13.0	54.3	32.6
rs4565946	G=>A	1.4	23.2	75.4	10.8	27.0	62.2

Generally there was a small shift towards recessive alleles in ASD kids, although it was not huge. Whether it contributes significantly to the condition is hard to say.

Serotonin Transporter SLC6A4

One SNP, the rs25531 SNP has been associated with lower levels of transport of serotonin.

SNP	Risk Allele	ASD +/+	ASD +/-	ASD -/-	Normal +/+	Normal +/-	Normal -/-
rs25531	C=>T	26/72	0/72	0/72	6/46	0/46	1/46

There is a higher level of expression of the TT allele in ASD individuals with lower expression of the short allele in normal individuals, period.

GAD Glutamic Acid Decarboxylase

Control of neuronal signaling can be achieved either through degradation or inactivation of the neurotransmitters, or through firing of inhibitory neurons using GABA. GABA synthesis is ultimately controlled by the enzyme Glutamic acid decarboxylase (GAD).

SNP	Risk Allele	ASD +/+	ASD +/-	ASD -/-	Normal +/+	Normal +/-	Normal -/-
rs2739023	A=>G	5.3	37.3	57.3	4.3	38.3	57.4
rs2241165	T=>C	6.3	37.5	56.3	4.3	42.6	53.2
rs2058725	T=>C	6.7	28.0	65.3	4.3	38.3	57.4
rs3791850	G=>A	6.8	24.7	68.5	4.3	38.3	57.4
rs769407	G=>C	16.2	28.4	55.4	4.3	48.9	46.8
rs3791851	T=>C	14.9	28.7	55.4	4.3	48.9	46.8
rs3838275	C=>T	13.3	48.0	38.7	17.0	46.8	36.2
rs12185692	C=>A	12.2	47.3	40.5	17.0	42.6	40.4
rs3791878	G=>T	12.2	35.1	52.7	12.8	34.0	53.2
rs10432420	G=>A	9.3	42.7	48.0	2.2	50.0	47.8
rs701492	C=>T	17.7	31.6	50.6	6.4	51.1	42.6
rs3791878	G=>T	12.3	34.2	53.4	12.8	34.0	53.2
rs1978340	G=>A	10.1	46.4	43.5	13.2	36.8	50.0
rs2236418	A=>G	4.1	26.0	69.9	0.0	20.0	80.0

For the majority of the SNPs there was little difference in the frequency of the minor alleles. The exceptions were rs769407, rs3791851, rs10432420 (all in GAD1) and rs2236418 (in GAD2). Given the distribution of the minor alleles, it would appear that there is normally some selective disadvantage for these SNPs in GAD.

COMT

Degradation/inactivation of the amino acid-based neurotransmitters occurs using COMT and MAO. Much has been made of COMT SNPs due to the requirement of COMT for SAM for activity.

SNP	Risk Allele	ASD +/+	ASD +/-	ASD -/-	Normal +/+	Normal +/-	Normal -/-
rs6269	A=>G	13.9	50.6	35.4	23.9	30.4	45.7
rs165599	A=>G	11.8	51.3	36.8	15.2	45.7	39.1
rs165722	T=>C	26.4	55.6	18.1	26.1	45.7	28.3
rs464612	T=>C	14.3	48.1	37.7	23.9	30.4	45.7
rs5993882	T=>G	5.6	38.9	55.6	4.3	36.2	59.6
rs480	A=>G	27.8	54.4	17.7	26.2	45.7	28.3
rs4633	T=>C	26.6	54.4	19.0	26.1	45.7	28.3
rs769224	G=>A	0.0	1.3	98.7	0.0	6.4	93.6
rs737866	T=>C	7.9	44.7	47.4	8.5	29.8	61.7
rs737865	A=>G	7.9	44.7	47.4	10.6	29.8	59.6
rs2239393	A=>G	15.8	50.0	34.2	24.4	28.9	46.7
rs174675	C=>T	10.0	44.3	45.7	6.5	41.3	52.2
rs1544325	G=>A	11.3	54.9	33.8	20.0	53.3	26.7
rs174696	T=>C	11.3	39.4	49.3	2.2	33.3	64.4
rs174699	T=>C	0	15.8	84.2	0.0	10.9	89.1
rs9332377	C=>T	1.4	26.0	72.6	2.8	27.8	69.4
rs165774	G=>A	7.0	43.7	49.3	11.1	31.1	57.8
rs5993883	T=>G	27.6	55.3	17.1	22.2	48.9	28.9
rs740601	T=>G	14.1	50.7	35.2	25.0	29.5	45.5

Raw data analysis of SNPs shows a moderate difference between ASD and normal individuals. Pooled data, though shows a different pattern, where may individuals had multiple +/+ alleles and others none. There seemed to be a shared association between rs165722, rs480 and rs4833. These 3 alleles had very similar frequencies both within genotypes and also between ASD and normal individuals, for the +/+ allele, however, the -/- condition appeared to have a lower than expected frequency, possibly suggesting selection against the SNP-type. Thus the frequency had dropped from 28% to 18%.

MAO

MAO represents one of the few X-linked genotypes examined. It is the second of the enzymes involved in inactivation/break-down of the neuor-amines, and is critically dependent upon vitamin B2 for function. Given that a male's phenotype is only dependent upon one copy of the gene, whereas the female required two copies we have compared male genotypes, and separately compared female genotypes.

Male Genotypes

SNP	Risk Allele	ASD +	ASD -	Normal +	Normal -
rs6323	G=>T	62.1	37.9	66.7	33.3
rs2072743	T=>C	54.7	45.3	58.6	41.4
rs979605	G=>A	37.1	62.9	38.5	61.5
rs1799836	T=>C	44.4	55.6	62.1	37.9
rs1799835	T	3.2	96.8	0.0	100
rs1800466	=>A	9.5	90.5	3.6	96.4
rs1465108	G=>A	32.8	67.2	28.6	71.4
rs5906883	C=>A	36.5	63.5	25.0	75.0
rs1137070	C=>T	40.6	59.4	27.6	72.4
rs3027452	G=>A	18.8	81.3	20.0	80.0
rs5905512	G=>A	41.3	58.7	33.3	66.7
rs3027407	G=>A	44.8	55.2	26.3	73.7

Female Genotypes

SNP	Risk Allele	ASD +	ASD +/-	ASD -/-	Normal +/+	Normal +/-	Normal -/-
rs6323	G=>T	36.4	45.5	18.2	54.5	39.4	6.1
rs2072743	T=>C	40	50.0	10.0	48.5	45.5	6.1
rs979605	G=>A	11.1	55.6	33.3	3.7	44.4	51.9
rs1799836	T=>C	10.0	60.0	30.0	27.3	51.5	21.2
rs1799835	T	0.00	0.0	100	0.0	0.0	100
rs1800466	=>A	0.0	0.0	100	0.0	0.0	100
rs1465108	G=>A	10.0	60.0	30.0	6.5	37.5	56.3
rs5906883	C=>A	9.1	54.5	36.4	6.1	39.4	54.5
rs1137070	C=>T	10.0	50.0	40.0	6.1	39.4	54.5
rs3027452	G=>A	0.0	40.0	60.0	0.0	33.3	66.7
rs5905512	G=>A	30.0	50.0	20.0	6.5	71.0	22.6
rs3027407	G=>A	11.1	55.6	33.3	7.1	32.1	60.7

rs3027452 AA conditionally lethal in females, however, not in males.

rs5905512 AA highly expressed in females with ASD rate is 4.6 times as high.

In both males and females, for rs5906883 and rs1137070 the frequency of the minor allele was increased.

For the majority of the SNPs there was very little difference between either ASD or normal individuals, nor was there difference between males and females.

Variation in vitamin B12 and folate processing proteins

Biochemically, ASD individuals are found to be very deficient in functional vitamin B2 and universally deficient in vitamin B12. The possibility exists, therefore, that genetic mutations in the enzymes involved in vitamin B12 processing and folate processing are partially responsible for the condition.

MTHFR

This enzyme is the "gateway" for moving folate cycling within the folate cycle which has been substituted to form 5,10-methylene-tetrahydrofolate out of the cycle and into the methylation cycle, whilst simultaneously reducing the methylene group to 5-methyltetrahydrofolate. The enzyme is critically dependent upon FAD (one of the two active forms of vitamin B2) and NADP (an active form of vitamin B3) for its activity.

SNP	Risk Allele	ASD +	ASD +/-	ASD -/-	Normal +/+	Normal +/-	Normal -/-
rs2066470	G=>A	1.4	17.3	81.3	0.0	19.1	80.9
rs1801131	T=>G	12.5	36.3	51.3	6.4	42.6	51.1
rs17367504	A=>G	3.8	26.3	70.0	2.1	27.7	70.2
rs1801133	G=>A	6.4	43.6	50.0	17.0	46.8	36.2
rs2274976	C=>T	1.3	13.9	84.8	0.0	12.8	87.2
rs1476413	C=>T	7.6	38.0	54.4	8.5	36.2	55.3
rs1703739	G=>A	4.0	25.3	70.7	0.0	26.1	73.9
rs3737964	C=>T	6.8	41.9	51.4	8.5	21.3	70.2
rs4846048	A=>G	12.3	42.5	45.2	8.7	26.1	65.2
rs4846049	G=>T	13.3	41.3	45.3	6.7	42.2	51.1

For the majority of the SNPs there was very little difference between ASD and normal individuals, particularly when one considers that for all but 2 of the SNPs the dominant allele predominated. One would not therefore expect that the SNP-type in MTHFR was causative for ASD.

MTR

The enzyme methionine synthase is involved in remethylating homocysteine to form methionine, and in the acquisition of the methyl group from 5MTHF to regenerate methyl B12. The activity of the enzyme is critically dependent upon methylB12 concentration.

SNP	Risk Allele	ASD +	ASD +/-	ASD -/-	Normal +/+	Normal +/-	Normal -/-
rs1805087	A=>G	2.5	40.5	57.0	2.1	23.4	74.5
rs2789352	C=>A	16.9	40.8	42.3	13.0	45.7	41.3
rs1092525	A=>G	2.8	43.7	53.5	2.2	23.9	73.9
rs2275568	C=>T	11.1	48.6	40.3	26.1	43.5	30.4
rs3820571	T=>G	22.1	33.8	44.1	16.3	46.5	37.2
rs1206026	G=>A	9.7	40.3	50.0	23.9	43.5	32.6
rs1206057	G=>C	9.7	38.9	51.4	26.1	41.3	32.6
rs3768142	T=>G	18.2	41.6	40.3	13.0	50.0	37.0
rs2275566	G=>T	12.3	40.0	47.7	13.6	47.7	38.6
rs2275565	G=>T	11.7	41.6	46.8	2.3	30.4	67.4
rs1092523	C=>T	18.1	43.1	38.9	13.0	47.8	39.1

There was some evidence of selection against the minor epitopes in MTR, namely rs2275568, rs1206026, and rs120605, which by frequency appear to be linked. Negative selection against this epitope could occur if the group represented a group that was selected against in B12 deficiency. Thus, as the methyl B12 concentration is critical for activity of MTR, individuals with lower affinity for vitamin B12 would be selected against in a low B12 environment. Theoretically, the offspring would have a more active enzyme, or at least one that had a higher affinity for methyl B12.

MTRR

The enzyme methionine synthase reductase is involved in remethylating inactive Co(II)B12, formed as an oxidation product of *Co(I)B12. The enzyme is critically dependent upon FMN/FAD and NADPH+ for activity.

SNP	Risk Allele	ASD +	ASD +/-	ASD -/-	Normal +/+	Normal +/-	Normal -/-
rs1801394	G=>A	27.8	54.4	17.7	10.6	48.9	40.4
rs162036	A=>G	1.3	38.8	60.0	0.0	12.8	87.2
rs11802059	G=>A	9.3	40.0	50.7	17.0	48.9	34.0
rs1532268	C=>T	8.9	41.8	49.4	17.0	48.9	34.0
rs3776467	A=>G	8.1	56.8	35.1	0.0	53.2	46.8
rs9332	G=>A	1.4	38.9	59.7	0.0	12.8	87.2

Interpreting SNP frequency can be a bit ambiguous due to two potentially conflicting reasons. First, the SNP may contribute to the condition that you are looking at. Such would be the case of rs1801394 and rs3776467, where there is shift from the more common GG/AA alleles to the less common, and presumably less active AA/GG alleles  (respectively). The data on rs1801394 is somewhat different to that reported by SNPedia, where the AA variant was much more common than the GG variant. However, the GG variant is much more common in the Indian Muslim population (4). This would be regarded as contributing to lower B12 levels. Second, the SNP may be actively selected against due to environment, that rs11802059 and rs1532268 have been selected away from the less common A and T alleles to the more common, and presumably more active G and C alleles, respectively, possibly due to environmental pressure in the womb to lower active B12 levels.

DHFR

The enzyme dihydrofolate reductase, is an enzyme in the folate cycle responsible for converting dihydrofolate to tetrayhydrofolate. The enzyme is critically dependent upon NADPH+ for activity.

SNP	Risk Allele	ASD +	ASD +/-	ASD -/-	Normal +/+	Normal +/-	Normal -/-
rs1643649	T=>C	12.0	37.3	50.7	14.9	38.3	46.8
rs1643659	T=>C	12.5	38.9	48.6	14.9	38.3	46.8
rs1677693	G=>T	12.5	38.9	48.6	14.9	38.3	46.8

There was a marginal negative shift in frequency away from the less active allele, however, it is unlikely that this would contribute much to the condition

Folate Receptor

Certain mutations within the folate receptor appear to be conditionally lethal, whilst some have been associated with a higher risk of Spina Bifida.

SNP	Risk Allele	ASD +	ASD +/-	ASD -/-	Normal +/+	Normal +/-	Normal -/-
rs2071010	G=>A	0.0	9.3	09.7	0.0	4.3	95.7
rs651933	G=>A	25.3	44.3	30.4	14.9	46.8	38.3
rs7925545	A=>G	1.3	2.5	96.3	0.0	8.5	91.5
rs7926875	C=>A	0.0	16.7	83.3	0.0	0.0	100.0

There was a significantly higher incidence of rs651933 in ASD, however examination of populatoins from countries across the world reveal that the alleles segregate normally in many populations. Thus, the impact of the allele is likely to be minimal. The extremely low incidence of the minor alleles in rs2071010, rs7925545, and rs8026875, suggest that these alleles may be conditionally lethal, as such the increase in the +/+ numbers for the 3 alleles may be associative for ASD.

MTHFDIL

Methylenetetrahydrofolate dehydrogenase (MTHFDIL) variants have been associated with an increased risk of heart defects..

SNP	Risk Allele	ASD +	ASD +/-	ASD -/-	Normal +/+	Normal +/-	Normal -/-
rs1076991	Y=>C	22.8	46.8	30.4	27.7	51.1	21.3
rs1801131	T=>G	12.5	36.3	51.3	6.4	42.6	51.1
RS2236224	G=>A	9.6	47.9	42.5	13.0	58.7	28.3

There was a small shift in relative frequencies of the 3 three SNPs tabulated, and similarly with rs11764661, rs17349743, rs6922269, rs803422, rs702465 adn rs7571842, however the shifts were small and so whilst possibly associative for ASD would not be regarded as causative in our estimation.

X-chromosome

The logical place to look for SNPs that have a greater effect on boys than on girls is the X chromosome. In boys there is an all or none chance of getting a "defective" SNP, whilst in girls the frequency of double SNPs would be one in four, hence the frequency of defective SNPs would be twice that of girls. If a two-site SNP approach is used, then the chance of getting 2 defective SNPs in boys would be four times the chance of receiving two sets of defective SNPs in girls. This then might explain the four-fold higher incidence of ASD in boys when compared to girls. To date such paired SNPs have not been identified (at least by us).

Others

Whilst some of the more obvious SNPs have not been found to be associative for ASD, there are a number of other SNPs in other genes that have been found to be more highly expressed in ASD individuals

rs6691117    Complement receptor A=>G frequency 0.39 ASD, 0.29 normal

rs2070045    Sortilin related receptor T=>G frequency 0.33 ASD, 0.15 normal

Summary

Examination of over 600 SNPs obtained from 180 individuals with autism and compared to persons with chronic fatigue syndrome (240 individuals) and "normal" individuals (100 individuals) has identified over 60 SNPs that are more highly expressed in the autistic individuals. The SNPs fall into linked clusters (i) vitamin D processing (6 SNPs), (ii) Fatty Acid Desaturase (5 SNPs), (iii) neurotransmitter processing (23 SNPs, COMT, GAD1, TPH1, TPH2, TH) (iv) no known function (15 SNPs). Of these, potentially the vitamin D processing SNPs could be regarded as causative for the condition, whilst the neurotransmitter processing SNPs could be regarded as affecting behaviour.

We have been unable to find any single SNP that is universal in kids with ASD. Using a cumulative analysis, however, we have found that in the associative SNPs for ASD that we have identified, autistic individuals generally have 10-20 ASD associative SNPs, whilst CFS individuals and normals have less than 8 associative SNPs. This would point to there being some genetic associations for the conditions, whether this is due to in-womb selection in a nutritionally deficient mother is open for discussion.