Systematically characterizing the roles of E3-ligase family members in inflammatory responses with massively parallel Perturb-seq

Geiger-Schuller, Kathryn; Eraslan, Basak; Kuksenko, Olena; Dey, Kushal Kumar; Jagadeesh, Karthik A.; Thakore, Pratiksha I.; Karayel, Özge; Yung, Andrea R.; Rajagopalan, Anugraha; Meireles, Ana M.; Yang, Karren; Amir-Zilberstein, Liat; Delorey, Toni; Phillips, Devan; Raychowdhury, Raktima; Moussion, Christine; Price, Alkes L.; Hacohen, Nir; Doench, John G.; Uhler, Caroline; Rozenblatt-Rosen, Orit; Regev, Aviv

doi:10.1101/2023.01.23.525198

Cited by 7 publications

(9 citation statements)

References 164 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We next asked if SC-VAE can capture how perturbations are related to each other such that learned perturbation map is biologically relevant [23]. To test this, we relied on the well-established assumption that perturbations in genes encoding members of the same protein complex should lead to similar phenotypes an thus more similar cell profiles [6, 8]. Accordingly, We adapted previous methodology, using the CORUM complex database [24] and the following process:…”

Section: Resultsmentioning

confidence: 99%

“…Some sources of technical noise are those common to all scRNA-seq studies, such as the sequencing depth and count over-dispersion, but others are specific to the perturbation context. For example, the capture rate of the guide RNA may be noisy, and therefore the association of a single cell to a perturbation may have some error [8]. Second, guide expression in a cell does not not necessarily mean that the cell has been genetically perturbation, because of varying efficiency rates of the Cas9 enzyme and the time delay from genetic perturbation to an effect on the target protein [9, 3].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A Supervised Contrastive Framework for Learning Disentangled Representations of Cell Perturbation Data

Tu,

Hütter,

Wang

et al. 2024

Preprint

Self Cite

View full text Add to dashboard Cite

CRISPR technology, combined with single-cell RNA-Seq, has opened the way to large scale pooled perturbation screens, allowing more systematic interrogations of gene functions in cells at scale. However, such Perturb-seq data poses many analysis challenges, due to its high-dimensionality, high level of technical noise, and variable Cas9 efficiency. The single-cell nature of the data also poses its own challenges, as we observe the heterogeneity of phenotypes in the unperturbed cells, along with the effect of the perturbations. All in all, these characteristics make it difficult to discern subtler effects. Existing tools, like mixscape and ContrastiveVI, provide partial solutions, but may oversimplify biological dynamics, or have low power to characterize perturbations with a smaller effect size. Here, we address these limitations by introducing the Supervised Contrastive Variational Autoencoder (SC- VAE). SC-VAE integrates guide RNA identity with gene expression data, ensuring a more discriminative analysis, and adopts the Hilbert-Schmidt Independence Criterion as a way to achieve disentangled representations, separating the heterogeneity in the control population from the effect of the perturbations. Evaluation on large-scale data sets highlights SC-VAE's superior sensitivity in identifying perturbation effects compared to ContrastiveVI, scVI and PCA. The perturbation embeddings better reflect known protein complexes (evaluated on CORUM), while its classifier offers promise in identifying assignment errors and cells escap- ing the perturbation phenotype. SC-VAE is readily applicable across diverse perturbation data sets.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

A Supervised Contrastive Framework for Learning Disentangled Representations of Cell Perturbation Data

Tu,

Hütter,

Wang

et al. 2024

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Single-cell screens link genetic perturbations to genome-wide transcriptional phenotypes. Previous studies 39,40 have identified shared gene programs that are commonly altered by multiple perturbations. If multiple distinct perturbations (defining a perturbation module) affect a common set of genes (defining a gene program), the rationale is that these perturbed genes/enhancers may contribute to breast cancer through shared mechanisms.…”

Section: Resultsmentioning

confidence: 99%

“…First, to identify perturbation modules that yield similar transcriptional phenotypes, we combined cells with a given regulatory element perturbed as a ‘meta-cell’ and clustered meta-cells by expression similarity 40 . We first filtered out lowly expressed genes and focused on the top 1,000 highly variable genes.…”

Section: Resultsmentioning

confidence: 99%

Enhancer regulatory networks globally connect non-coding breast cancer loci to cancer genes

Wang,

Armendariz,

Wang

et al. 2023

Preprint

View full text Add to dashboard Cite

Genetic studies have associated thousands of enhancers with breast cancer. However, the vast majority have not been functionally characterized. Thus, it remains unclear how variant-associated enhancers contribute to cancer. Here, we perform single-cell CRISPRi screens of 3,512 regulatory elements associated with breast cancer to measure the impact of these regions on transcriptional phenotypes. Analysis of >500,000 single-cell transcriptomes in two breast cancer cell lines shows that perturbation of variant-associated enhancers disrupts breast cancer gene programs. We observe variant-associated enhancers that directly or indirectly regulate the expression of cancer genes. We also find one-to-multiple and multiple-to-one network motifs where enhancers indirectly regulate cancer genes. Notably, multiple variant-associated enhancers indirectly regulate TP53. Comparative studies illustrate sub-type specific functions between enhancers in ER+ and ER- cells. Finally, we developed the pySpade package to facilitate analysis of single-cell enhancer screens. Overall, we demonstrate that enhancers form regulatory networks that link cancer genes in the genome, providing a more comprehensive understanding of the contribution of enhancers to breast cancer development.

show abstract

“…Intuitively, a physical interaction between two proteins suggest that they might participate in a shared biomolecular pathway or complex. Thus, perturbations of genes coding for physically interacting proteins might lead to similar effects [40]. For each perturbation x, ϕ x is a node embedding of x in the PPI network, such as node2vec [41].…”

Section: Methodsmentioning

confidence: 99%

Sequential Optimal Experimental Design of Perturbation Screens Guided by Multi-modal Priors

Huang,

Lopez,

Hütter

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

Understanding a cell’s expression response to genetic perturbations helps to address important challenges in biology and medicine, including the function of gene circuits, discovery of therapeutic targets and cell reprogramming and engineering. In recent years, Perturb-seq, pooled genetic screens with single cell RNA-seq (scRNA-seq) readouts, has emerged as a common method to collect such data. However, irrespective of technological advances, because combinations of gene perturbations can have unpredictable, non-additive effects, the number of experimental configurations far exceeds experimental capacity, and for certain cases, the number of available cells. While recent machine learning models, trained on existing Perturb-seq data sets, can predict perturbation outcomes with some degree of accuracy, they are currently limited by sub-optimal training set selection and the small number of cell contexts of training data, leading to poor predictions for unexplored parts of perturbation space. As biologists deploy Perturb-seq across diverse biological systems, there is an enormous need for algorithms to guide iterative experiments while exploring the large space of possible perturbations and their combinations. Here, we propose a sequential approach for designing Perturb-seq experiments that uses the model to strategically select the most informative perturbations at each step for subsequent experiments. This enables a significantly more efficient exploration of the perturbation space, while predicting the effect of the rest of the unseen perturbations with high-fidelity. Analysis of a previous large-scale Perturb-seq experiment reveals that our setting is severely restricted by the number of examples and rounds, falling into a non-conventional active learning regime called “active learning on a budget”. Motivated by this insight, we develop IterPert, a novel active learning method that exploits rich and multi-modal prior knowledge in order to efficiently guide the selection of subsequent perturbations. Using prior knowledge for this task is novel, and crucial for successful active learning on a budget. We validate IterPertusing insilico benchmarking of active learning, constructed from a large-scale CRISPRi Perturb-seq data set. We find that IterPertoutperforms other active learning strategies by reaching comparable accuracy at only a third of the number of perturbations profiled as the next best method. Overall, our results demonstrate the potential of sequentially designing perturbation screens through IterPert.

show abstract

Systematically characterizing the roles of E3-ligase family members in inflammatory responses with massively parallel Perturb-seq

Cited by 7 publications

References 164 publications

A Supervised Contrastive Framework for Learning Disentangled Representations of Cell Perturbation Data

A Supervised Contrastive Framework for Learning Disentangled Representations of Cell Perturbation Data

Enhancer regulatory networks globally connect non-coding breast cancer loci to cancer genes

Sequential Optimal Experimental Design of Perturbation Screens Guided by Multi-modal Priors

Contact Info

Product

Resources

About