The 10 alleles that were found to be correlated to educational attainment and which I used for my landmark study on selection for intelligence (Mankind Quarterly, 2013) have been replicated on a sample of children: http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0100248#pone-0100248-g002

Or they didn't want to risk Bonferroni's wrath

The 10 alleles that were found to be correlated to educational attainment and which I used for my landmark study on selection for intelligence (Mankind Quarterly, 2013) have been replicated on a sample of children: http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0100248#pone-0100248-g002

Recently, there was another replication in a Han sample.

Zhu, B., Chen, C., Moyzis, R. K., Dong, Q., & Lin, C. (2015). Educational attainment-related loci identified by GWAS are associated with select personality traits and mathematics and language abilities. Personality and Individual Differences, 72, 96-100.

A recent genome-wide association study of educational attainment identified three significant single nucleotide polymorphisms (SNPs) (rs9320913, rs11584700, and rs4851266)....Association analysis for individual SNPs showed significant associations between rs4851266 and a measure of language ability (Chinese word recognition), and between rs12202969 and a personality trait (fear of negative evaluation) and a measure of mathematical ability (number paired-associates learning). A polygenic score based on these three SNPs was also significantly associated with the measures of mathematical and language abilities.

Another gene (CNV not SNP, so not present in 23andMe raw data): http://www.ncbi.nlm.nih.gov/pubmed/25287832

Intriguing. Potentially big finding.

Apparently it increases the risk of autism too: http://www.plosgenetics.org/article/info%3Adoi%2F10.1371%2Fjournal.pgen.1004241

Intriguing. Potentially big finding.

From: Volkmar Weiss [mailto:Volkmar-Weiss@t-online.de]
Sent: Mittwoch, 21. Januar 2015 10:59
To: 'James.Sikela@ucdenver.edu'
Subject: DUF1120 copy number is linearly associated with IQ

Dear Professor Sikela:

Yesterday I read in Human Genetics 134 (2015) 67-75 the full text of your publication on „DUF1220 copy number is linearly associated with increased cognitive function as measured by total IQ and mathematical aptitude scores”,

see http://www.ncbi.nlm.nih.gov/pubmed/25287832 ,

full text: https://drive.google.com/file/d/0B3c4TxciNeJZcHg4ek9nWDZGcWs/view?pli=1

As I did catch sight of your Fig. 2 on page 72 I got deeply impressed, deeply moved, deeply: A linear association between copy arrayCGH based CON2 copy ratio versus WISC IQ, extending in the IQ range between 80 and 140. This is a breakthrough, a centennial breakthrough! My congratulation to your lab and the cooperating colleagues in New Zealand!

We are aware: What you have discovered is the tip of an iceberg. But it is the tip!

Your were even clever enough to obtain a patent for the determination of IQ by this copy number variation (CPV),

see http://www.google.com/patents/WO2014028768A2?cl=en

DUF means a protein “domain of unknown function“, containing a number of genes, especially of the NBPF family, each of it highly polymorphic.

We need and this research may already under way in your lab:

1. Family studies of the inheritance of CON2 copy ratio and IQ. There all over the world thousands of families with more than one gifted in the IQ range around 130, ready to be probands.

2. Representative population data.

Until now it was assumed by the majority that IQ differences are caused by thousands of genetic polymorphisms each of it making a small contribution of plus or minus 1, 2 or even3 IQ points. Therefore the environment must play an important or decisive role. Since decades, whoever hinted that this thousand-genes-theory did not agree with the data of segregation of IQ within families, see for example http://www.v-weiss.de/majgenes.html , was seen as an obstinate crank. However, in 1972, already in my dissertation I wrote that the hypothetical major gene locus of general intelligence could turn out to be a series of alleles. And in 1992: “Of course, the allele M2 could also be understood as an abstraction and be in reality a series of n alleles with small differences; but with a large difference to the M1 allele or an allele-1 series.”


The difference between the means of the hypothetical M1M1 and M2M2 is about 30 IQ points. This is the range, what you found! The other hundreds of polygenes which, of course, influence mental power under certain circumstances may add up to IQ differences of 20 points in extreme and rare cases, but because the minor genes are segregating independently of each other, their effects as a sum are normally distributed making only a plus or minus of about 5 IQ-points in the general population.


As we know, in the search for major effects on IQ all genome-wide association studies (GWAS) were a failure. Therefore, the conclusion had to be drawn that the explanation had to be found in previously unexplored regions of the genome.

Therefore, since some years I suggest to look for copy number variations and the application of homozygosity array mapping within families of the highly gifted. See my monograph “Die Intelligenz und ihre Feinde” (Intelligence and its Enemies). Graz 2012, page 236 to CPV: „Da es sehr gut vorstellbar ist, daß diese Art der genetischen Variabilität auch in der Genetik des normalen IQ eine wichtige Rolle spielt, konzentrieren sich die Hoffnungen gegenwärtig auf weitere Erforschung dieser ‚Copy number variations‘ (CPV).“

You and your lab had the knowledge. You did it. My congratulation.


I wish you the possibility, the freedom and the courage to extend your findings. You will need it.


Sincerely yours

Volkmar Weiss

www.v-weiss.de



http://www.collective-evolution.com/2014/10/13/junk-dna-or-does-it-reveal-the-source-of-our-mind-intelligence/

Sample sizes are tiny for the Duff1220 CNV study. Won't that be a problem?

These found nothing, using bigger samples. However they don't seem to use the same method/tool as in the DUFF1220 study.

n 800, only including rare CNVs:
http://www.ncbi.nlm.nih.gov/pubmed/23417127

n 723. Including both common and rare CNVs. Also the sample is from New Zealand:

http://www.ncbi.nlm.nih.gov/pubmed/23383111

These studies did not include the DUF1220 protein domain.

By the way, DUF1220 is not a gene, "Domain of Unknown Function" = DUF. This protein domain cooperates with at least 20 genes, especially of the NBPF family..

These studies did not include the DUF1220 protein domain.

By the way, DUF1220 is not a gene, "Domain of Unknown Function" = DUF. This protein domain cooperates with at least 20 genes, especially of the NBPF family..

Ah ok.

Did they correlate the number of copy number variations of a set of genes to IQ or was it the specific genes within the domain? As in was it the burden of copy number variation that linearly associated with one of the samples or was it the specific genetic variants themselves? Or was it both?

I recommend to read the original paper.

The amount of DUF1220 of the clade CON2 is highly correlated with IQ. The range of variation is more than 2 SD of IQ! Such a range for the effects of the major gene was predicted by Volkmar Weiss, see "Major genes of general intelligence" (1992).

According to the classic definition of a gene it is entity which segregates according to the Mendelian rules. This definition holds also for SNPs and copy numbers.

As I know from well-informed insiders, all over the world several research groups try to replicate and to extend Sikela's findings with great numbers of probands.