I am delighted to see on BioRxiv.org the publication from my PhD student Jack Humphrey. Jack works in collaboration between the UCL Genetics Institute and the UCL Institute of Neurology.
Jack's work, during the first year of his PhD, has focused on understanding the molecular mechanisms behind devastating neurodegenerative disorders like ALS and FTD. These neurodegenerative disorders are raising interesting questions for computational biologists. Many of the variants- mostly rare- associated with these disorders implicate genes that play a role in RNA processing, for example RNA binding proteins. However, this is far from the whole story and many questions remain to precise the molecular mechanisms that initiate the disease process.
Back in August 2015, Ling et al published in Science an exciting paper showing that when a gene called TDP-43 is not expressed at the usual level, small sections of the genomes become unexpectedly expressed, as if new uncontrolled exons were appearing in the genome. TDP-43 is implicated in ALS and FTD in two ways. Firstly, rare variants in that gene are associated with familial forms of ALS-FTD, which provide a reliable causal relationship between that gene and disease. Secondly, TDP-43 inclusions, i.e. accumulation of that protein, are almost systematically found in the post mortem brains of ALS and FTD patients.
This Ling et al paper is obviously interesting because this cryptic exon observation could provide a mechanism linking TDP-43 and disease. Not taking anything away from that publication, the results are not highly quantitative in the sense that the authors did not make full use of the biological replicates, a pillar of quantitative biology. We felt that this publication opened many interesting questions that were worth investigating using a computational approach, which would provide a good starting point for Jack's PhD.
The BioRxiv publication looks at available TDP-43 datasets and mine these data to better quantify the cryptic exon mechanism. Jack's analysis confirms that the impaired competitive RNA binding between TDP-43 and other splicing factors is the most likely mechanism leading to cryptic exons. In addition, Jack shows that the gene containing these cryptic exons also are on average down-regulated, most likely because of non-sense mediated decay induced by these cryptic exons. This down-regulation suggests that the impact of cryptic exons on the regulatory machinery is potentially large, strengthening the view that this specific mechanism is a plausible candidate for pathogenesis.
Taken together, these results confirm many of the findings of Ling et al and provide new directions. We plan to take these results to the next stage and see how that signal can be detected in other neuro-degenerative disease models, not directly TDP-43 related. Most importantly, we will be looking at human data to assess whether similar signals can be identified. However, the challenges in human post mortem brains are much greater, mostly because of the much more degraded RNA and the extent in neuronal loss in late stage patients. Time will tell how far we will be able to go with these questions.