Es Castell

So, being able to query SF regulation systematically using single-cell perturbation screens revealed a model of their regulation during carcinogenesis:
- oncogenic lesions activate MYC
- activation of oncogenic-like and inactivation tumor suppressor-like SFs through their cross-regulation

We then asked which pathways could be linking carcinogenic mutations with SF regulation.

Among the candidates, MYC stands out as the top one!

MYC had already been shown to be a regulator of SFs, but it had never been confirmed unbiasedly.

Now we can look at how each Perturb-seq KD changes the activity of cancer splicing programs.

Perturbations that activate one program inactivate the other, and those causing carcinogenic regulation are enriched in SFs themselves.

Cross-regulation among SFs drives their carcinogenic regulation.

While directly porting our method to GE did not fully recapitulate the regulation of SFs observed with exon inclusion, adjusting gene-based SF activity with a single fully-connected NN layer did the job!

Actually, we confirmed the coordinated regulation of SF programs in melanoma carcinogenesis!

How can we study the regulation of these cancer SF programs systematically?

Single-cell perturbation screens have great throughput but don’t have exon inclusion resolution to estimate SF activity with our approach.

So we need to estimate SF activity from GE to analyze Perturb-seq data.

We previously showed that changes in target exon inclusion are highly informative of the activity of SFs: doi.org/10.1101/2024...

We identified two pan-cancer SF programs that show coordinated regulation during carcinogenesis: oncogenic SFs become more active than tumor suppressor-like SFs.

For example, running the notebooks using the WT and the MUT sequence of LMNA causing limb girdle muscular dystrophy 1B ( doi.org/10.1038/nrg.... ), Pangolin predicts a decrease in exon 9's 5' splice site strength:

Ciutadella, 12/2024

Low tie in Fanø