Spectra decomposes gene expression data into interpretable programs using prior knowledge.
Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser . In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.
The minimization of cell-type influence allows Spectra to identify factors that are shared across cell types. We show that Spectra outperforms existing approaches and solves longstanding challenges in tumor immune contexts, including the identification of an interpretable tumor reactivity factor in CD8T cells and a new invasion program in macrophages, which associate with response and resistance to cancer immunotherapy, respectively.
In contrast to Spectra, Slalom’s accuracy drops substantially as the number of active gene sets increases . Moreover, Slalom can only assess a few dozen gene sets before run time becomes prohibitive, whereas Spectra scales to hundreds of thousands of cells and hundreds of gene programs. When run on a graphics processing unit , Spectra outperforms all methods, including NMF and the GPU-based expiMap .
The common simplifying assumption made by factor analysis methods is that factors combine linearly to drive expression, which is not always the case. Uncovering interpretable nonlinear relationships is a future goal of factorization methods development. To avoid penalizing novel factors that have no relation to the annotations, we introduce a weighting matrix that scales the computation of gene–gene similarity scores by factor-specific weights that are learned from the data. Factors that have low weight are not used in computing edge probabilities, whereas factors with high weights influence the edge probabilities directly.
A second advantage of the graph prior is its scalability. Although gene sets may be highly overlapping, especially when curated from several separate databases, this redundancy is eliminated when storing information at the level of gene–gene relationships. Redundant gene sets will be merged into highly overlapping communities, and so two redundant gene sets can be approximately described by a single factor.
$$-\mathop{\sum }\limits_{i=1}^{{n}_{c}}\mathop{\sum }\limits_{j=1}^{p}\left[{\alpha }_{c,i,:K}^{\top }{\theta }_{j}+{\alpha }_{c,i,K+1:}^{\top }{\theta }_{cj}\right]$$\}=\frac{{\bar{L}}_{c}-{\bar{L}}_{c}}{{\bar{L}}_{c}}\)but computed per cell type to represent cell-type-specific information content. Specifically, given a marker list associated with a factor and with$${C}_{c,k}=\mathop{\sum }\limits_{m=2}^{M}\mathop{\sum }\limits_{l=1}^{m-1}\log \frac{{D}_{c}},{g}_{l}^{})+1}{{D}_{c}})}$$.
South Africa Latest News, South Africa Headlines
Similar News:You can also read news stories similar to this one that we have collected from other news sources.
Ex-DA Seth Williams has a new story to tell now that his federal supervised release is overSeth Williams, pleading guilty six years ago, said he “squandered” the trust of the city, his friends and family. Now he says the feds ganged up on him with the Catholic Church.
Read more »
Inter Miami keeps playoff hopes alive while testing fan loyalty amid Lionel Messi injury sagaFORT LAUDERDALE, Fla. — Kamal Miller sat glued to rain-soaked grass as unfulfilled fans scurried off into the South Florida night.
Read more »
Inter Miami keeps playoff hopes alive while testing fan loyalty amid Lionel Messi injury sagaFORT LAUDERDALE, Fla. — Kamal Miller sat glued to rain-soaked grass as unfulfilled fans scurried off into the South Florida night.
Read more »