Apologies for my blogging absence, everyone. I’m working towards a rapidly-approaching book deadline and much of my writing energies have gone towards that project. However, rest assured, my blogging will become more regular once again. In the meantime, I’d like to talk about some of our upcoming work.
Over the last year, my collaborators and I have been working on a genetics project investigating major genotype-phenotype correlations in intellectual disability. Although this project is not quite complete as it’s currently under review at a journal, we elected to make the draft available through the preprint server, bioRxiv. Therefore, since the meat of the project is already available online, I’m going to review some of our findings. However, please keep in mind these data have not yet passed the peer review process and should therefore be taken with a grain of salt.
The current project is a continuation of an early study we published in 2016, reporting significant patterns of functional enrichment within high-risk autism genes associated with intellectual disability. To summarize the major theme of that study, we found that high-risk autism genes typically function as epigenetic regulators, controlling the timing and expression of other genes during development. This has been reported in other studies, however they didn’t limit their investigations to genes with high autism penetrance. While there are other gene types associated with autism risk, mutations in epigenetic regulators seem to have the highest penetrance for autism (i.e., they lead to the autism phenotype most often).
We’ve continued to study these same genes in the current project, grouping them according to additional clinical phenotypes to see if any genotype-phenotype clustering occurs. The major clinical phenotypes we looked at (in addition to the presence/absence of autism and epilepsy diagnoses) were:
- complex (3+) facial malformations
- simple (1-2) facial malformations
- neurodegenerative-like features, consisting of brain atrophy and various combinations of movement disorders
Facial malformations and neurodegenerative-like features were selected because they occurred frequently in the conditions we studied, and because they overlapped inconsistently, suggesting that they were due to different underlying biological processes. When they do overlap in a single syndrome, it is instead suggestive of genetic pleiotropy– meaning that a single gene mutation can have multiple different effects.
So, we divided our large gene list according to each gene’s associated features: autism, epilepsy, facial malformations, and neurodegenerative-like features. Those different clinical features gave us 18 subgroups and we ran each of these subgroups through GeneMANIA to derive a list of additional interacting genes. Finally, we built one large network of genes using Cytoscape, which showed us to what extent these 18 groups overlapped one another. Some subgroups clearly divided into even smaller groups, suggesting they didn’t form singular clusters and therefore probably overlapped little etiologically.
Rectangle = intellectual disability with autism (with/without epilepsy); triangle = epilepsy (without autism) with intellectual disability; octagon = intellectual disability without autism or epilepsy; maroon node = complex facial malformations; green node = neurodegenerative-like features; purple node = complex facial malformations and neurodegenerative-like features; light blue node = simple facial malformations; red node = simple facial malformations and neurodegenerative-like features; gold node = none; aqua line = co-expression; gold line = genetic interaction; purple line = physical interaction.
However, a number of our main subgroups of interest did form relatively nonoverlapping clusters, suggesting that the phenotypes we studied are reflective of unique molecular interaction networks and that syndromic autism does indeed form major genotypic and phenotypic clusters. In the above image, the autism groups with complex facial malformations (maroon rectangle), neurodegenerative-like features (green rectangle), and a mix of the two (purple rectangle) form three distinctive clusters that share overlap specifically via the genes, FMR1 and MECP2, both of which are of major interest in autism research.
We’re currently starting work to extend these data, re-test their reliability, and determine whether they have tissue-specific expressions that could be useful in determining how and why complex facial malformations and neurodegenerative-like features associate with syndromic autism and whether specific molecular networks are responsible for these associations. We do all this in the hopes of better understanding and defining the underlying biology of certain autism clusters. At least now we have a better idea that clusters likely do exist in autism. We’ll see if this extends beyond syndromic forms of the condition as well…