Sparse dictionary learning recovers pleiotropy from human cell fitness screens

Pan, Joshua; Kwon, Jason J.; Talamas, Jessica A.; Borah, Ashir A.; Vázquez, Francisca; Boehm, Jesse S.; Tsherniak, Aviad; Žitnik, Marinka; McFarland, James M.; Hahn, William C.

doi:10.1016/j.cels.2021.12.005

Cited by 20 publications

(31 citation statements)

References 94 publications

(180 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Thus, INTS10-13-14-C7orf26 represents a functionally and biochemically distinct module of Integrator, consistent with concurrent studies ( Figure 3H ; Funk et al, 2021 ; Pan et al, 2022 ; Wainberg et al, 2021 ). Although Integrator has been subjected to extensive structural analyses, it has been difficult to resolve the INTS10-13-14 components in relation to the rest of the complex, and inclusion of C7orf26 may facilitate future structural efforts.…”

Section: Resultssupporting

confidence: 89%

“…Overexpression of C7orf26 with INTS10, INTS13, and INTS14 enabled the purification of a stable INTS10-13-14-C7orf26 complex by size-exclusion chromatography ( Figure 3E ; Data S2 [ Figures i and ii ]). We also detected an interaction between the Drosophila C7orf26 ortholog and fly Integrator and observed co-essentiality between C7orf26 and INTS10, INTS13, and INTS14 in the Cancer Dependency Map ( Data S2 [ Figure iii ]) ( Pan et al, 2022 ; Wainberg et al, 2021 ). These results suggest that C7orf26 is a core subunit of an INTS10-13-14-C7orf26 Integrator module.…”

Section: Resultsmentioning

confidence: 75%

See 1 more Smart Citation

Mapping information-rich genotype-phenotype landscapes with genome-scale Perturb-seq

Replogle

Saunders

Pogson

et al. 2022

Cell

197

220

View full text Add to dashboard Cite

Section: Resultssupporting

confidence: 89%

Section: Resultsmentioning

confidence: 75%

Mapping information-rich genotype-phenotype landscapes with genome-scale Perturb-seq

Replogle

Saunders

Pogson

et al. 2022

Cell

197

220

View full text Add to dashboard Cite

“…This dataset was constructed by performing CRISPR-Cas9 knockouts on an immortalized human cell line (RPE1-hTERT) and subjecting each knockout variant to 31 genotoxic stressors. We show that the core processes described by our SSD decomposition are enriched for particular gene annotations and compare our decomposition to one identified by Webster [17].…”

Section: Applications To Empirical Datamentioning

confidence: 82%

“…First, we evaluate whether the core processes described by our solution are enriched for [19] and curated by [17]. (b) The locus and phenotype rotation test indicate there is both locus and phenotype sparsity.…”

Section: B Robustness Of Gene Knockouts To Genotoxins In Human Cell L...mentioning

confidence: 99%

Inferring sparse structure in genotype-phenotype maps

Petti

Reddy

Desai

2022

Preprint

View full text Add to dashboard Cite

Phenotypic variation across related individuals is often correlated: a high or low value of one phenotype tends to be associated with a high or low value of others. This may reflect lower-dimensional structure in the genotype-phenotype map, such that genotype affects a relatively small set of unobserved "core" processes that in turn determine the observed phenotypes. Identifying low-dimensional structure in high-dimensional genotype-phenotype data thus offers the promise of inferring the identity and genetic basis of core biological processes, as well as the way in which core processes determine each observed phenotype. However, inferring this lower-dimensional structure requires appropriate biologically motivated constraints, even with high-throughput genotype-phenotype measurements. Here, we show that several recent empirical genotype-phenotype data sets exhibit evidence of sparse structure, and that a sparsity-favoring matrix decomposition approach can accurately recover latent processes if each genetic perturbation affects few core processes or if each phenotype is affected by few core processes. Motivated by this, we develop a generally applicable framework based on penalized matrix decomposition for sparse structure discovery (SSD) and apply it to three empirical datasets spanning adaptive mutations in yeast, genotoxin robustness assay in human cell lines, and genetic loci identified from a yeast cross. More generally, we propose sparsity as a guiding prior for resolving latent structure in empirical genotype-phenotype maps.

show abstract

“…Though genes which operate in the same biological process or pathway have similar knockout fitness profiles, genes whose functional interaction is context-dependent can have their profile similarities weakened by inclusion of cell lines or contexts where the interaction is not present. We (Kim et al, 2022) and others (Pan et al, 2022) have recently shown how analysis of CRISPR genetic screen data can reveal proteins that have multiple roles in a cell, leading to functional interactions with mutually exclusive partners. We extend the Pearson correlation network by using partial correlation to identify these “moonlighting” genes, whose interactions with one gene are boosted when the effect of a second gene is removed (and vice versa).…”

Section: Discussionmentioning

confidence: 99%

Optimal construction of a functional interaction network from pooled library CRISPR fitness screens

Gheorghe

Hart

2022

Preprint

View full text Add to dashboard Cite

Functional interaction networks, where edges connect genes likely to operate in the same biological process or pathway, can be inferred from CRISPR knockout screens in cancer cell lines. Genes with similar knockout fitness profiles across a sufficiently diverse set of cell line screens are likely to be co-functional, and these coessentiality networks are increasingly powerful predictors of gene function and biological modularity. While several such networks have been published, most use different algorithms for each step of the network construction process. In this study, we identify an optimal measure of functional interaction and test all combinations of options at each step -- essentiality scoring, sample variance and covariance normalization, and similarity measurement -- to identify best practices for generating a functional interaction network from CRISPR knockout data. We show that Bayes Factor and Ceres scores give the best results, that Ceres outperforms the newer Chronos scoring scheme, and that covariance normalization is a critical step in network construction. We further show that Pearson correlation, mathematically identical to ordinary least squares after covariance normalization, can be extended by using partial correlation to detect and amplify signals from moonlighting proteins which show context-dependent interaction with different partners.

show abstract

Sparse dictionary learning recovers pleiotropy from human cell fitness screens

Cited by 20 publications

References 94 publications

Mapping information-rich genotype-phenotype landscapes with genome-scale Perturb-seq

Mapping information-rich genotype-phenotype landscapes with genome-scale Perturb-seq

Inferring sparse structure in genotype-phenotype maps

Optimal construction of a functional interaction network from pooled library CRISPR fitness screens

Contact Info

Product

Resources

About