Genes for loneliness, health club attendance, bar hopping and churchgoing, all in one study!

I am attempting to critique this study:

Elucidating the genetic basis of social interaction and isolation (Day et al.)

There are many directions I can go with such a critique.  The most appealing and easiest, would be to mock it with a couple of quick quotes and be done with it.  Then, I think to myself, are there people out there that take a study like this seriously?  And, of course there are a lot of people who take a study like this seriously.  

I'm hoping, though, that there are a few scientists who have been holding onto these GWAS studies as some sort of proof of all kinds of mental constructs, who might have a bit of a crisis of confidence when reading something like this.  Perhaps they would like to dismiss this as an outlier, or misguided in some way.

Here's where they have a problem.  Because, this study was done by the book.  It has all the elements used to prove that there are genetic associations for these traits, just as studies are done to find associations for IQ, mental disorders, and personality traits.  So if you are touting GWAS studies related to any of these traits, you need to explain why your study is good and this one is ridiculous, or you need to embrace this ridiculous study.  There is no in between.

With that in mind, I will go through how this study follows the same formula as your cherished studies and you can decide which side of the health club attendance gene fence you sit on.
Let's start with their primary result:

In up to 452,302 UK Biobank study participants, we perform genome-wide association study analyses for loneliness and regular participation in social activities. We identify 15 genomic loci (P < 5 × 10−8) for loneliness, and demonstrate a likely causal association between adiposity and increased susceptibility to loneliness and depressive symptoms. 

Again, with the UK Biobank, you can invent any trait in your mind, as long as it can be elucidated in a questionnaire,  and then proceed to find genetic loci for this "trait."  Do you like horror movies?  Let's find a gene for that.  Do you think raisins are tasty?  Must be a gene.  Is blue your favorite color? blue gene...  etc.  So the first question we have to ask, is how many genomic loci we would expect at random to be (P < 5 × 10−8) ?  In other words, are these just random false positives?  Like any GWAS study, this one does not address this in any direct way.  I will once again point out my own method for determining this and invite any of the researchers performing these studies to perform my suggested test.
Secondly, if you want to try and make a case that fat people are lonelier, then you don't need to look at unreplicated genetic associations for the two traits for comparison.  You can just look at that directly.

Let's look at some of the bonus traits that were looked at: gym rats, bar hoppers, and churchgoers:

Further loci were identified for regular attendance at a sports club or gym (N = 6 loci), pub or social club (N = 13) or religious group (N = 18). 

Interesting that there is no apparent crossover amongst these three different sets of loci.  Apparently, our desire to be parts of groups are specific enough, that there are different sets of genes involved in each activity.  Makes one wonder how many social activities (existing or conceived of) we could name before we get a match, or would we just run out of genes?
So how do they "prove" these are valid?  They use a common method I see in these GWAS studies, which I have heard repeatedly argued as a basis for the validity of genetic associations without the need for replication:

Across these traits there was strong enrichment for genes expressed in brain regions that control emotional expression and behaviour. We demonstrate aetiological mechanisms specific to each trait, in addition to identifying loci that are pleiotropic across multiple complex traits.

In other words, they claim that the genes from these loci express themselves in the brain, so that proves they are on the right track.   I have seen this used to show that IQ genes, depression genes, schizophrenia genes, etc. are valid.  The fact that these genetic loci rarely if ever replicate does not seem to give pause.  I will point out a possible reason for this:  This is an example of wishful connections.  Again, we have no random controls.  If we did, let's say, hand this group of scientists a set of loci that were found to be significantly associated with diabetes without telling them it wasn't from the current study,  would they also find "strong enrichment for genes expressed in brain regions that control emotional expression and behaviour"?  I'm guessing they would.  In any case, if you were buying into this kind of analysis before, does it start to waver a bit?  I hope so, if you have any kind of critical eye.

Can I point out a bit of circular logic from this paper?

For example, does the onset of depressive symptoms and cognitive decline cause withdrawal from social engagement, or is this a consequence of the discrepancy between preferred and experienced social relations. To address this question, and provide insights into the potential biological mechanisms that contribute to loneliness, we identify associated genetic variants in the UK Biobank study

In other words, since we found all these genes, the answer is the latter.  Never mind that these genetic associations have never been replicated and not even demonstrated to be anything other than false positives.

The next part of the study is more of the same.  However, I include it, because these kinds of convoluted connections are used in so many GWAS studies to find a way to justify the validity of genetic associations that theywill never replicate.

By integrating gene expression and epigenetic data we sought to identify the relevant cell/tissue types implicated in the regulation of loneliness. We observed enrichment of association signals in regions surrounding genes that are preferentially expressed in several brain tissues (e.g., cerebellum, basal ganglia, and cortex; Fig. 2 Supplementary Table 1), in addition to enrichment for several epigenetic marks also in the basal ganglia, cortex and foetal brain (Supplementary Data 2). We next used FUSION9 to identify individual genes implicated by associated eQTL effects in GTEx brain tissues (Supplementary Data 3). Across 9178 transcripts in 9 tissue types, we identified 8 gene transcripts with expression levels putatively linked-to susceptibility to loneliness (GPX1, C1QTNF4, C17orf58, MTCH2, BPTF, RP11-159N11.4, CRHR1-IT1 and PLEKHM1). BPTF encodes a transcription regulator that is highly expressed in foetal brain and is implicated in neurodegenerative diseases. GPX1 and MTCH2 are implicated in multiple metabolic pathways, including mitochondrial function.

Again, the simple question to ask here is, "Compared to what"?  Where is your control?  How hard is it to find CNS correlations to basically any genetic loci you find in a GWAS study?

The next part of the study is also a common component of GWAS studies:  Finding other unreplicated loci from other traits that we also see in our current study.  Let's start with this one:

Furthermore, a recently reported signal for risk-taking propensity11showed far stronger association with sports/gym attendance (CADM2-rs7627971, P = 5.8 × 10−10) than pub/social club (P = 3.0 × 10−3) or religious group (P = 3.9 × 10−2) attendance.

Leaving aside the fact that there is little to indicate that working out in a gym is more of a risk than hanging out in bars,  this is only one genetic association from one study,  which has never been replicated and is about as ridiculous as this study (See my critique of the "risk-taking" study here).
They go through a series of dubious associations between their traits and various genetic associations (my favorite being "age at first sexual intercourse").  I'll let the reader wade through this if they have an interest.

Oh, I almost forgot:  Polygenic Risk Score.  This statistical abomination is now being used to justify unreplicated genetic associations and they make no bones about it:

A limitation of our analysis was the absence of comparably sized independent replication studies to replicate associations with individual loci. This represents a challenge for genetic studies of complex traits with extremely large discovery datasets, such as UK Biobank, particularly for traits that are uncommonly measured. However, cumulative assessment of the polygenic risk score for loneliness in an unrelated sample that demonstrates the overall validity of our study design and analytical approach.

Here is what they claim related to this:

A genetic risk score comprised of the 15 lead SNPs predicted loneliness in an independent set of 7556 individuals (P = 0.025). 

I don't see where the genetic risk score is derived, nor what is meant by "predicted loneliness" .  I would be interested in how these scores are derived and have yet to see them posted anywhere and I invite the authors to shed light on it.

In conclusion, if you are a behavioral genetic scientist, you have a GWAS study before you  that finds loci for a trait, finds a correlation to CNS, finds connections to other traits and uses polygenic risk score to bolster their claims.  I ask you, do you think that there are genetic associations for loneliness, bar hopping, gym attendance, and church going?  I think you either have to question the whole GWAS phenomenon, or you have to embrace these findings.  I leave the choice to you...