Integrated Analysis of Gene Expression Differences in Twins Discordant for Disease and Binary Phenotypes, 2018, Patel and Tangirala

[Andy]

It claims to have looked at twins with one twin with CFS but talks about Chronic Fatigue, so how useful to us this could be is questionable.

Abstract

While both genes and environment contribute to phenotype, deciphering environmental contributions to phenotype is a challenge. Furthermore, elucidating how different phenotypes may share similar environmental etiologies also is challenging. One way to identify environmental influences is through a discordant monozygotic (MZ) twin study design. Here, we assessed differential gene expression in MZ discordant twin pairs (affected vs. non-affected) for seven phenotypes, including chronic fatigue syndrome, obesity, ulcerative colitis, major depressive disorder, intermittent allergic rhinitis, physical activity, and intelligence quotient, comparing the spectrum of genes differentially expressed across seven phenotypes individually.
Second, we performed meta-analysis for each gene to identify commonalities and differences in gene expression signatures between the seven phenotypes.

In our integrative analyses, we found that there may be a common gene expression signature (with small effect sizes) across the phenotypes; however, differences between phenotypes with respect to differentially expressed genes were more prominently featured. Therefore, defining common environmentally induced pathways in phenotypes remains elusive. We make our work accessible by providing a new database (DiscTwinExprDB:http://apps.chiragjpgroup.org/disctwinexprdb/) for investigators to study non-genotypic influence on gene expression.

Open access at https://www.nature.com/articles/s41598-017-18585-3

 

[Valentijn]

It's using data from a "chronic fatigue" study which combined CF and CFS patients. That one found no differences, so it's a bit odd that this newer study (which also looks at data from twin studies in other diseases) did. It looks like they did correct for making multiple comparisons, but used a method for it (False Discovery Rate) which is useful for identifying potentially relevant trivial differences, but subject to a higher rate of false positives.

The actual genes involved can be seen at http://apps.chiragjpgroup.org/disctwinexprdb/ . Based on the small "meandiff" values between the CF and non-CF twins, I think that means the effect sizes are very small? But the genes with statistically significant differences in expression are SEC62 (endoplasmic reticulum), RPL22 (ribosomes), NACA (ribosomes and endoplasmic reticulum), SUGT1 (centromeres), CREM (cAMP), EIF1 (mRNA translation), ACTG1 (muscle cytoskeleton), ACTB (non-muscle cytoskeleton), SEMA6D (dorsal root ganglion signaling), and DNM1L (mitochondrial & peroxisomal division).

Note that this study is about gene expression, and is looking at the products produced by genes rather than the genetic code itself. So SNPs and mutations are not involved in this research, though it is possible it influenced the gene expression - that is unlikely, however, since identical twins were used.

 

[Amw66]

It's using data from a "chronic fatigue" study which combined CF and CFS patients. That one found no differences, so it's a bit odd that this newer study (which also looks at data from twin studies in other diseases) did. It looks like they did correct for making multiple comparisons, but used a method for it (False Discovery Rate) which is useful for identifying potentially relevant trivial differences, but subject to a higher rate of false positives.

The actual genes involved can be seen at http://apps.chiragjpgroup.org/disctwinexprdb/ . Based on the small "meandiff" values between the CF and non-CF twins, I think that means the effect sizes are very small? But the genes with statistically significant differences in expression are SEC62 (endoplasmic reticulum), RPL22 (ribosomes), NACA (ribosomes and endoplasmic reticulum), SUGT1 (centromeres), CREM (cAMP), EIF1 (mRNA translation), ACTG1 (muscle cytoskeleton), ACTB (non-muscle cytoskeleton), SEMA6D (dorsal root ganglion signaling), and DNM1L (mitochondrial & peroxisomal division).

Note that this study is about gene expression, and is looking at the products produced by genes rather than the genetic code itself. So SNPs and mutations are not involved in this research, though it is possible it influenced the gene expression - that is unlikely, however, since identical twins were used.

Can prolonged stress, environmental exposure and potentially viral infection not also alter gene expression?

 

[Snow Leopard]

It's using data from a "chronic fatigue" study which combined CF and CFS patients. That one found no differences, so it's a bit odd that this newer study (which also looks at data from twin studies in other diseases) did.

Was that the one where they're just like "there were no differences, data not shown" or something to that effect?

 

[Valentijn]

Can prolonged stress, environmental exposure and potentially viral infection not also alter gene expression?

Potentially, though I'm not sure how rigorous research has been in the area. It's been pretty horrible when involving large amounts of SNPs, due to failure to correct for making multiple comparisons, and gene expression studies could easily have the same problem. It could also be difficult to make a connection between a specific cause and any effect seen.

One of the benefits of any twin study (identical or fraternal) is that life experiences are usually very similar, so different exposure to stress, toxins, infections, etc also don't seem very likely, though definitely more likely than different DNA. I think it's more likely that the differences in gene expression (if truly significant) are the result of the diseases, rather than the cause of them.

Was that the one where they're just like "there were no differences, data not shown" or something to that effect?

Not sure ... I usually don't pay much attention to CF-not-CFS studies

 

[Jonathan Edwards]

This looks to me like a complete confusion in the minds of the authors. They say:

One way to identify environmental influences is through a discordant monozygotic (MZ) twin study design.

That is not the case because monozygotic twins can be discordant for a disease because of stochastic causal factors.

All you demonstrate by showing differences in gene EXPRESSION between affected and unaffected twins is that one is ill and the other is not. Epidemiology gives a pretty good indication as to whether an illness is genetic, environmental or stochastic (or mixed). In the vast majority of situations there is little need to study gene expression in monozygotic twins because gene expression studies are studies of what happens once an illness has started, not studies of what causes an illness. If there is no epidemiological indication of the problem being genetic it is vanishingly unlikely that any differences found will be due to actual gene mutations.

This sort of research seems to have completely lost sight of our basic knowledge of tissue structure and physiology. The human body is treated just as a bag of genes, much as a word processing programme might be treated as a string of bits. If you do not even know how those bits are used for word processing you are unlikely to work out where a bug in the programme is situated.

Edit: agree with what Valentijn has said and I have probably duplicated.