Genetic identification of brain cell types underlying schizophrenia, Skene, 2018

[ME/CFS Science Blog]

Abstract​

With few exceptions, the marked advances in knowledge about the genetic basis of schizophrenia have not converged on findings that can be confidently used for precise experimental modeling. By applying knowledge of the cellular taxonomy of the brain from single-cell RNA sequencing, we evaluated whether the genomic loci implicated in schizophrenia map onto specific brain cell types. We found that the common-variant genomic results consistently mapped to pyramidal cells, medium spiny neurons (MSNs) and certain interneurons, but far less consistently to embryonic, progenitor or glial cells. These enrichments were due to sets of genes that were specifically expressed in each of these cell types. We also found that many of the diverse gene sets previously associated with schizophrenia (genes involved in synaptic function, those encoding mRNAs that interact with FMRP, antipsychotic targets, etc.) generally implicated the same brain cell types. Our results suggest a parsimonious explanation: the common-variant genetic results for schizophrenia point at a limited set of neurons, and the gene sets point to the same cells. The genetic risk associated with MSNs did not overlap with that of glutamatergic pyramidal cells and interneurons, suggesting that different cell types have biologically distinct roles in schizophrenia.

Link: https://www.nature.com/articles/s41588-018-0129-5

 

[ME/CFS Science Blog]

Another cell type study that I wanted to test the DecodeME data to.

They used an older Mouse Brain Atlas from the Linnarson lab at the Karolinska Institute (the same group that published the human brain atlas). The data from different dissections was spread out over multiple publications such as the ones below:

Redirecting

https://www.science.org/doi/10.1126/science.aaa1934

This paper (Skene et al.) combined these into one dataset with 9,970 cells with 24 level 1 and 149 level 2 cell types.

It's an older and not so big dataset but it has been applied to multiple other conditions such as in the paper by Olislagers et al. 2021. So that makes it interesting to take a look at.

Comprehensive analyses of RNA-seq and genome-wide data point to enrichment of neuronal cell type subsets in neuropsychiatric disorders - Molecular Psychiatry

Neurological and psychiatric disorders, including substance use disorders, share a range of symptoms, which could be the result of shared genetic background. Many genetic loci have been identified for these disorders using genome-wide association studies, but conclusive evidence about cell types...

www.nature.com

 

[ME/CFS Science Blog]

The data is available in the R package MAGMA.celltyping by the authors.
neurogenomics/MAGMA_Celltyping: Find causal cell-types underlying complex trait genetics

Mouse gene have been mapped to human equivalents by them. They also changed the gene expression specificity into 40 quantile (I presume because the data was too skewed and LDSC and MAGMA work with linear regressions). I followed their approach and used MAGMA because that's what I know best. One issue is that the data doesn't include IDs for the genes, only names, so I used the NCBI37.3.gene.loc file to match those.

The results for DecodeME using the level 1 cell types looks like this:



The top level 2 cell types looked like this:

 

[ME/CFS Science Blog]

There is one caveat to this analysis, the olisalgers et al. 2021 paper found that this linear MAGMA analysis with quantiles performed worse (did not align with their other methods). They prefer LDSC or MAGMA with the top 10% of genes as binary annotation. They also restrict their GWAS to the approximately 1.2 million hapmap3 SNPs and I found that this had quite a large influence on the results.




I also tried LDSC but I'm not confident that I did it right because it is quite complex to build annotation files for each combination of cell type x chromosome. I found that some of the top cell types were the same but that no cell type was significant.

 

[ME/CFS Science Blog]

These analyses were a bit above my level as an amateur so don't take them too seriously.

 

[Felis Catus]

Mouse gene have been mapped to human equivalents by them.

Does that mean that the approach is based on the assumption that orthologs between human and mouse have conserved function and expression patterns?