She observed that the knocking down RPL10 and RPL24 has the same effect as knocking down RPL5 and RPL11, known members of the ribosomal stress pathway, showing that her fingerprinting-identified candidates were indeed real!

Next, she identified two proteins that were likely involved in the ribosomal stress response from their transcriptome profile. We used FlexPlex to knock down those genes with siRNA, treated them with Nutlin-3a, oligohashed them and then used custom probes to confirm the loss of the intended target.

From there she determined that only under certain doses were these inhibitors HDAC-specific and above a certain threshold, they then act like pan-HDAC inhibitors

First, she was curious about the specificity of HDAC inhibitors that are purported to be HDAC subclass specific versus pan-HDAC inhibitors, so we applied FlexPlex with oligo-based hashing to simultaneously observe the effect of these compounds on the transcriptome and histone modifications

We also identified new candidates for involvement in the integrated stress response, including RACK1, showing an increase in ATF4 levels upon perturbation

One observation of interest was a set of genes that seemingly responded to feedback from mTOR signaling - going up with both Ragulator and TSC loss and going down with mTORC1 and GATOR1 loss - sets of complexes that have opposite effects on mTOR signaling

Finally, we clustered perturbations from the validation FlexPlex assay into different categories based on their transcriptome and we observed how the different perturbations converge on mTOR signaling

We validated 120 of these hits with another FlexPlex assay, observing pRPS6 changes in most of them.

However, we wanted to bring these predictions genome-wide - so we applied the model to the K562 GWPS dataset from the Weissman Lab, observing 586 perturbations in that dataset that we predicted to significantly alter pRPS6

From there we modeled pRPS6 activity based on transcriptomic profile using RIDGE regression and applied it to other datasets from with known or inferred changes in pRPS6 - including datasets in completely different cell types including organoids!

We then curated that list to 50 of the top perturbations producing higher or lower pRPS6 and ran a FlexPlex assay with that, producing a “glossary” dataset with paired transcriptome-pRPS6 profiles for each perturbation.

We then applied FlexPlex to understand regulators of mTORC1 signaling. We first did a genome-wide bulk CRISPRi screen on pRPS6 in K562 cells to provide a candidate list of relevant perturbations which, while noisy, reconstructed the known mTOR pathway and revealed new potential regulators.

We demonstrate the effectiveness of FlexPlex through perturbing known components of the mTOR pathway, observing increases in pRPS6 in cells with TSC1 knocked down and decreases in cells with RPTOR knocked down. We confirmed excellent guide assignment and robust transcriptome profiling with high UMIs

There is a ton of IF showing the remarkable pathology of both the knockout organoid astrocytes (left) and astrocytes from resected patient cortical tuber tissue

The single cell sequencing was performed at multiple important timepoints in early development and in multiple genetic backgrounds, all with our meticulously engineered conditional two-hit TSC2 knockout alleles and knockout reporter cassettes

There are significant changes from the initial pre-print posted .... some time ago.... including profiling retinal organoids, adding new benchmarks and demonstrating an additional method to incorporate transcriptomic data