Scientists discover 17 genetic variations that raise risk of depression

[Bob]

Scientists discover 17 genetic variations that raise risk of depression
Research involved 300,000 people of European ancestry and could lead to better understanding of the condition
https://www.theguardian.com/science...-of-depression?CMP=twt_a-science_b-gdnscience

To unpick possible genetic associations, the researchers examined data collected by the consumer genetic testing company 23andme. Of the 300,000 people studied, 75,607 self-reported a clinical diagnosis of depression or were receiving treatment for the condition.

By comparing the frequency of common genetic variations in the people with depression compared to those without, the scientists discovered two genomic regions associated with the condition, one of which has previously been linked to epilepsy and intellectual disability.

Further analysis, including 23andme data from another 150,000 individuals, as well as clinical data from a worldwide multi-institutional collaboration involving nearly 19,000 individuals, threw up more results, with scientists identifying 17 genetic variants in 15 genome regions associated with depression. The findings are published in the journal Nature Genetics.

 

[Deltrus]

Full PDF: http://docdro.id/sz5Orel

Why does it not list which variant is bad?

 

[Valentijn]

Only 5 of the 17 SNPs are actually on genes. The rest are between genes. Effect sizes are extremely tiny - even if they are accurate, they would have so little effect that it might as well be none at all.

Why does it not list which variant is bad?

In the supplementary excel file, the 2nd allele listed is the one they're looking at the effect of. In that file, the alleles are listed in alphabetical order, as they are in 23andMe data.

For the 5 SNPs on genes:

• rs2179744 - "A" is the risky allele
• rs301806 - "C" is the risky allele
• rs10786831 - "A" is the risky allele, though more common than the other allele (ie, most people would have this supposed risk factor)
• rs12552 - "A" is the risky allele
• rs6476606 - "A" is the risky allele

All of these "risky" alleles are extremely common, with between 35% and 55% allele frequency. That means between 58% and 80% of all people will have each SNP.

 

[cfs6691]

Only 5 of the 17 SNPs are actually on genes. The rest are between genes. Effect sizes are extremely tiny - even if they are accurate, they would have so little effect that it might as well be none at all.


In the supplementary excel file, the 2nd allele listed is the one they're looking at the effect of. In that file, the alleles are listed in alphabetical order, as they are in 23andMe data.

For the 5 SNPs on genes:

• rs2179744 - "A" is the risky allele
• rs301806 - "C" is the risky allele
• rs10786831 - "A" is the risky allele, though more common than the other allele (ie, most people would have this supposed risk factor)
• rs12552 - "A" is the risky allele
• rs6476606 - "A" is the risky allele

All of these "risky" alleles are extremely common, with between 35% and 55% allele frequency. That means between 58% and 80% of all people will have each SNP.

I haven't read any research into genetics but I have read a little bit about neurotransmitters.When scientists look at genes do they know which neurotransmitter or which receptor is affected and how when someone has a polymorphism?The SNPs between genes is it clear whether they play a role or not(if not why do they mention it)?

 

[Valentijn]

When scientists look at genes do they know which neurotransmitter or which receptor is affected and how when someone has a polymorphism?

Yes, when looking at a gene, researchers almost always know which enzyme it creates. Basically each enzyme is composed of a unique string of amino acids, and the alleles in the exons of each gene can be directly "read" to determine which known protein is created by that gene.

There are locations suspected of being genes, but with an unknown product. But those unknown gene locations are pretty easy to spot with computer software, due to repetitions of certain patterns which occur at the ends of genes. So it's very unlikely that any of the SNPs which are not on genes will turn out to be on undiscovered genes.

The SNPs between genes is it clear whether they play a role or not(if not why do they mention it)?

They don't play a role. But sometimes alleles which are fairly close together will also be inherited together. So it's possible that the SNPs which are not on genes reflect relevant but untested SNPs on nearby genes. But in this study a lot of those distances between the SNPs and the nearest genes are pretty huge, which makes it less likely that they're relevant to the functioning of that gene.

And even when close to the gene, at best it suggests that there might be something relevant on the gene. That would warrant a closer look at the gene in the future, not a conclusion of impact based on SNPs which aren't on the gene.