Highlights d Rare coordinated high expression states in cancer cells can drive therapy resistance d Gene networks with transcriptional bursting recapitulate these transcriptional states d Networks with low connectivity favorably give rise to these states d Parameters affecting transcriptional bursting are critical to produce these states

One of the most interesting, difficult, and potentially useful topics in computational biology is the inference of gene regulatory networks (GRNs) from expression data. Although researchers have been working on this topic for more than a decade and much progress has been made, it remains an unsolved problem and even the most sophisticated inference algorithms are far from perfect.In this paper, we review the latest developments in network inference, including state-of-the-art algorithms like PIDC, Phixer, and more. We also discuss unsolved computational challenges, including the optimal combination of algorithms, integration of multiple data sources, and pseudo-temporal ordering of static expression data. Lastly, we discuss some exciting applications of network inference in cancer research, and provide a list of useful software tools for researchers hoping to conduct their own network inference analyses.

Cellular memory describes the length of time a particular transcriptional state exists at the single-cell level. Transcriptional states with memory can underlie important processes in biology, including therapy resistance in cancer. Here we present a new experimental and computational approach for identifying gene expression states with memory at single-cell resolution by combining single-cell RNA sequencing (scRNA-seq) with cellular barcoding. With this technique, we can systematically quantify the full expression profile of these memory states as well as their population dynamics including the relative growth rates of cells within different states and the rates of switching between states. We applied this approach to human melanoma cells and uncovered memory gene expression states that are predictive of which cells will be resistant to combination BRAF and MEK inhibition. While these cells have not been treated with targeted therapies, they already express markers of resistance, and thus are referred to here as “primed” for resistance. From the scRNA-seq data alone, the drug-susceptible and primed cells appear as two distinct and unrelated cell populations. From the lineage barcodes, we found that most cells remain within the same state, demonstrating memory of gene expression. However, we also directly observe state switching as we find that about 18% of lineages contain cells that have switched states. While the molecular drivers of state switching are not immediately apparent from scRNA-seq data alone, we specifically analyzed lineages that undergo state switching and identified TGF-β and PI3K as drivers of state switching at the single-cell level. Through functional validation and single-cell RNA-sequencing, we show mechanistic single-cell manipulation of state switching that ultimately yields changes in cellular phenotype. We leverage these mechanisms of state switching to delay resistance by demonstrating that reducing the number of primed cells with a PI3K inhibitor prior to BRAF and MEK inhibition ultimately leads to fewer resistant colonies. Our results show that modulation of cellular signaling can globally regulate gene expression programs to overcome therapy resistance.

Mechanistic models of how single cells respond to different perturbations can help integrate disparate big data sets or predict response to varied drug combinations. However, the construction and simulation of such models have proved challenging. Here, we developed a python-based model creation and simulation pipeline that converts a few structured text files into an SBML standard and is high-performance- and cloud-computing ready. We applied this pipeline to our large-scale, mechanistic pan-cancer signaling model (named SPARCED) and demonstrate it by adding an IFNγ pathway submodel. We then investigated whether a putative crosstalk mechanism could be consistent with experimental observations from the LINCS MCF10A Data Cube that IFNγ acts as an anti-proliferative factor. The analyses suggested this observation can be explained by IFNγ-induced SOCS1 sequestering activated EGF receptors. This work forms a foundational recipe for increased mechanistic model-based data integration on a single-cell level, an important building block for clinically-predictive mechanistic models.

In isogenic cell populations, cells can switch back and forth between different gene expression states. These expression states can be biologically relevant. For example, a certain expression state may cause a tumor cell to be resistant to treatment, while another state may leave it vulnerable to treatment. However, estimating the rates of state-switching can be difficult, because experimentally measuring a cell's transcriptome often involves destroying the cell, so it can only be measured once. In this paper, we propose a computational method to estimate the rate of switching between expression states, given data from a Luria-Delbrück style fluctuation test that is experimentally simple and feasible. We then benchmark this method using simulated data to test its efficacy, with varying assumptions made about cell cycle timing distribution in the simulations.

One challenge in gene network inference is distinguishing between direct and indirect regulation. Some algorithms, including ARACNE and Phixer, approach this problem by using pruning methods to eliminate redundant edges in an attempt to explain the observed data with the simplest possible network structure. However, we hypothesize that there may be a cost in accuracy to simplifying the predicted networks in this way, especially due to the prevalence of redundant connections, such as feed forward loops, in gene networks. In this paper, we evaluate the pruning methods of ARACNE and Phixer, and score their accuracy using receiver operating characteristic curves and precision-recall curves. Our results suggest that while pruning can be useful in some situations, it may have a negative effect on overall accuracy that has not been previously studied. Researchers should be aware of both the advantages and disadvantages of pruning when inferring networks, in order to choose the best inference strategy for their experimental context.

In the noisy cellular environment, RNAs and proteins are subject to considerable stochastic fluctuations in copy numbers over time. As a consequence, single cells within the same isoclonal population can differ in their expression profile and reside in different phenotypic states. The dynamic nature of this intercellular variation, where individual cells can transition between different states over time makes it a particularly hard phenomenon to characterize. Here we propose a novel fluctuation-test approach to infer the kinetics of transitions between cell states. More specifically, single cells are randomly drawn from the population and grown into cell colonies. After growth for a fixed number of generations, the number of cells residing in different states is assayed for each colony. In a simple system with reversible switching between two cell states, our analysis shows that the extent of colony-to-colony fluctuations in the fraction of cells in a given state is monotonically related to the switching kinetics. Several closed-form formulas for inferring the switching rates from experimentally quantified fluctuations are presented. We further extend this approach to multiple cell states where harnessing fluctuation signatures can reveal both the topology and the rates of cell-state switching. In summary, our analysis provides a powerful approach for dissecting cell-state transitions based on a single time point measurement. This is especially important for scenarios where a measurement involves killing the cell (for example, performing single-cell RNA-seq or assaying whether a microbial/cancer cell is in a drug-sensitive or drug-tolerant state), and hence the state of the same cell cannot be measured at different time points.

No abstract