In silico RNA isoform screening to identify potential cancer driver exons with therapeutic applications.

Alternative splicing is crucial for cancer progression and can be targeted pharmacologically, yet identifying driver exons genome-wide remains challenging. We propose identifying such exons by associating statistically gene-level cancer dependencies from knockdown viability screens with splicing profiles and gene expression. Our models predict the effects of splicing perturbations on cell proliferation from transcriptomic data, enabling in silico RNA screening and prioritizing targets for splicing-based therapies. We identified 1,073 exons impacting cell proliferation, many from genes not previously linked to cancer. Experimental validation confirms their influence on proliferation, especially in highly proliferative cancer cell lines. Integrating pharmacological screens with splicing dependencies highlights the potential driver exons affecting drug sensitivity. Our models also allow predicting treatment outcomes from tumor transcriptomes, suggesting applications in precision oncology. This study presents an approach to identifying cancer driver exon and their therapeutic potential, emphasizing alternative splicing as a cancer target.

Disentangling the splicing factor programs underlying complex molecular phenotypes.

The regulation of exon inclusion through alternative splicing tunes the cell's behavior by increasing the functional diversity of the transcriptome and the proteome. Splicing factors work in concert to generate gene isoform pools that contribute to cell phenotypes yet their activity is controlled by multiple regulatory and signaling layers. This hinders identification of functional, phenotype-specific splicing factors using traditional single-omic measurements, such as their mutational state or expression. To address this challenge, we propose repurposing the virtual inference of protein activity by enriched regulon analysis (VIPER) to measure splicing factor activity solely from their downstream exon transcriptomic inclusion signatures. This approach is effective in assessing the effect of co-occurring splicing factor perturbations, as well as their post-translational regulation. As proof of concept, we dissect recurrent splicing factor programs underlying tumorigenesis including aberrantly activated factors acting as oncogenes and inactivated ones acting as tumor suppressors, which are undetectable by more conventional methodologies. Activation and inactivation of these cancer splicing programs effectively stratifies overall survival, as well as cancer hallmarks such as proliferation and immune evasion. Altogether, repurposing network-based inference of protein activity for splicing factor networks distills common, functionally relevant splicing programs in otherwise heterogeneous molecular contexts.

Reverse engineering neuron type-specific and type-orthogonal splicing-regulatory networks using single-cell transcriptomes.

Cell type-specific alternative splicing (AS) enables differential gene isoform expression between diverse neuron types with distinct identities and functions. Current studies linking individual RNA-binding proteins (RBPs) to AS in a few neuron types underscore the need for holistic modeling. Here, we use network reverse engineering to derive a map of the neuron type-specific AS regulatory landscape from 133 mouse neocortical cell types defined by single-cell transcriptomes. This approach reliably inferred the regulons of 350 RBPs and their cell type-specific activities. Our analysis revealed driving factors delineating neuronal identities, among which we validated Elavl2 as a key RBP for MGE-specific splicing in GABAergic interneurons using an in vitro ESC differentiation system. We also identified a module of exons and candidate regulators specific for long- and short-projection neurons across multiple neuronal classes. This study provides a resource for elucidating splicing regulatory programs that drive neuronal molecular diversity, including those that do not align with gene expression-based classifications.

Genes enriched in A/T-ending codons are co-regulated and conserved across mammals.

Codon usage influences gene expression distinctly depending on the cell context. Yet, the importance of codon bias in the simultaneous turnover of specific groups of protein-coding genes remains to be investigated. Here, we find that genes enriched in A/T-ending codons are expressed more coordinately in general and across tissues and development than those enriched in G/C-ending codons. tRNA abundance measurements indicate that this coordination is linked to the expression changes of tRNA isoacceptors reading A/T-ending codons. Genes with similar codon composition are more likely to be part of the same protein complex, especially for genes with A/T-ending codons. The codon preferences of genes with A/T-ending codons are conserved among mammals and other vertebrates. We suggest that this orchestration contributes to tissue-specific and ontogenetic-specific expression, which can facilitate, for instance, timely protein complex formation.

robustica: customizable robust independent component analysis.

BACKGROUND: Independent Component Analysis (ICA) allows the dissection of omic datasets into modules that help to interpret global molecular signatures. The inherent randomness of this algorithm can be overcome by clustering many iterations of ICA together to obtain robust components. Existing algorithms for robust ICA are dependent on the choice of clustering method and on computing a potentially biased and large Pearson distance matrix. RESULTS: We present robustica, a Python-based package to compute robust independent components with a fully customizable clustering algorithm and distance metric. Here, we exploited its customizability to revisit and optimize robust ICA systematically. Of the 6 popular clustering algorithms considered, DBSCAN performed the best at clustering independent components across ICA iterations. To enable using Euclidean distances, we created a subroutine that infers and corrects the components' signs across ICA iterations. Our subroutine increased the resolution, robustness, and computational efficiency of the algorithm. Finally, we show the applicability of robustica by dissecting over 500 tumor samples from low-grade glioma (LGG) patients, where we define two new gene expression modules with key modulators of tumor progression upon IDH1 and TP53 mutagenesis. CONCLUSION: robustica brings precise, efficient, and customizable robust ICA into the Python toolbox. Through its customizability, we explored how different clustering algorithms and distance metrics can further optimize robust ICA. Then, we showcased how robustica can be used to discover gene modules associated with combinations of features of biological interest. Taken together, given the broad applicability of ICA for omic data analysis, we envision robustica will facilitate the seamless computation and integration of robust independent components in large pipelines.

Chromosome segregation fidelity requires microtubule polyglutamylation by the cancer downregulated enzyme TTLL11.

Regulation of microtubule (MT) dynamics is key for mitotic spindle assembly and faithful chromosome segregation. Here we show that polyglutamylation, a still understudied post-translational modification of spindle MTs, is essential to define their dynamics within the range required for error-free chromosome segregation. We identify TTLL11 as an enzyme driving MT polyglutamylation in mitosis and show that reducing TTLL11 levels in human cells or zebrafish embryos compromises chromosome segregation fidelity and impairs early embryonic development. Our data reveal a mechanism to ensure genome stability in normal cells that is compromised in cancer cells that systematically downregulate TTLL11. Our data suggest a direct link between MT dynamics regulation, MT polyglutamylation and two salient features of tumour cells, aneuploidy and chromosome instability (CIN).

Combining CRISPRi and metabolomics for functional annotation of compound libraries.

Molecular profiling of small molecules offers invaluable insights into the function of compounds and allows for hypothesis generation about small-molecule direct targets and secondary effects. However, current profiling methods are limited in either the number of measurable parameters or throughput. Here we developed a multiplexed, unbiased framework that, by linking genetic to drug-induced changes in nearly a thousand metabolites, allows for high-throughput functional annotation of compound libraries in Escherichia coli. First, we generated a reference map of metabolic changes from CRISPR interference (CRISPRi) with 352 genes in all major essential biological processes. Next, on the basis of the comparison of genetic changes with 1,342 drug-induced metabolic changes, we made de novo predictions of compound functionality and revealed antibacterials with unconventional modes of action (MoAs). We show that our framework, combining dynamic gene silencing with metabolomics, can be adapted as a general strategy for comprehensive high-throughput analysis of compound functionality from bacteria to human cell lines.