Cancer Drivers: The Big Fish of Cellular Gangs
Cancer research is a lot like detective work. Like the gangster film private eye who must unravel a criminal scheme one back-alley deal at a time, scientific investigators are working to solve the cancer puzzle piece-by-piece. Identifying cancer drivers—the key factors fueling disease progression—is like the chase to identify the mastermind of a crime before it’s too late.
In these films, the “big fish”—the crime bosses at the center of it all—orchestrate their unlawful enterprises from the shadows, using a vast network of subordinates to do their bidding while staying off the radar of law enforcement. Tracing the criminal activity back to its mastermind requires teasing apart the web of associates linking the big fish to the crime.
Similarly, cancer drivers may only be discoverable through the implication of their associates. After over a decade of studying cancer, I’ve come to see these drivers as the hidden “crime bosses” of the disease."
Why Cancer Drivers Are Hard to Identify
Cancer arises from disruptions in gene regulation, the process governing which genes get turned “on” (expressed) or “off” (silenced), and when. Gene regulation is a complex process often involving multiple interrelated mechanisms that operate at various levels, including:
- Epigenetic modifications: Changes to DNA or histones, such as DNA methylation and histone acetylation, that affect gene accessibility.
- Transcriptional control: Regulation of gene expression by transcription factors and enhancers.
- Post-transcriptional regulation: Processes like alternative splicing or microRNA activity that modify RNA after transcription.
- Post-translational modifications: Changes to proteins, such as phosphorylation or ubiquitination, that alter their function or stability.
The multifarious nature of gene regulation can make it challenging to pinpoint what triggers and sustains a particular cancer, in part because many of the existing technologies for measuring gene expression only capture a single aspect of the process (for instance, transcriptional activity) at a time. Operating these technologies is expensive and time-consuming, forcing many researchers to employ a limited number of methods rather than the entire arsenal. These limitations, combined with the fact that many drivers may only be evident at certain levels of gene expression, makes identifying the culprit into a game of hide-and-seek. For instance, a researcher who relies on gene expression profiling to identify candidate drivers will be blind to drivers whose expression levels are unchanged but that are altered by, for example, post-translational modifications. Trying to identify cancer drivers solely through gene expression profiling is like trying to find the big fish at the scene of the crime—an often fruitless endeavor despite their central role and culpability.
Gene Networks: a Roadmap to Cancer Driver Detection
Until these technologies, as well as multi-omics approaches, become faster and more affordable, we are left with the challenge of identifying cancer drivers using a limited number of measurements. In gangster films, the big fish are often identified on the basis their associations with individuals more directly implicated in the crimes. A mastermind might arouse the suspicions of police for no reason other than their myriad connections to existing suspects. The key to identifying these central figures lies in their networks.
We wondered if the same principle might apply to cancer: Can we identify the drivers responsible for cancer and other diseases based on their networks?
The answer is yes! In fact, scientists have been using this approach in other fields for decades. For example, enzymes are well-established master regulators of metabolic pathways as well as drivers of many metabolic diseases. Measuring enzymes directly (through their expression) does not adequately capture enzymatic activity; to get a fuller, researchers instead rely on measurements of the enzymes’ substrates. In other words, they leverage the network to uncover the truth. Likewise, a holistic approach that considers cancer drivers within the context of their greater networks can provide deeper insights than focusing on individual genes or proteins alone.
scMINER: A Computational Detective
With this in mind, we developed scMINER, a computational tool that reconstructs gene networks to identify the "big fish" driving cancer and other serious diseases. scMINER leverages the data obtained through a single technology, singe-cell RNA sequencing, allowing researchers to gain a comprehensive picture of gene regulation from a single set of measurements.
scMINER is an all-in-one solution for analyzing single-cell transcriptomics data (see figure below). It offers unparalleled performance in three key functions:
- Unsupervised Cell Clustering: scMINER distinguishes closely related cell populations more accurately than five state-of-the-art algorithms.
- Network Inference: we compared scMINER to three established methods and found that it did a better job reconstructing gene networks.
- Hidden Driver Identification: Besides identifying key transcription factor drivers more accurately than the popular method SCENIC, scMINER is also able to identify the comparatively more abundant signaling protein drivers, providing a more comprehensive picture.
While scMINER can function as a standalone tool for analyzing single-cell RNA-Seq data, we have also developed a portal that offers additional visualization options, allowing users to explore their results interactively as well as share their findings with others. The scMINER Portal can be accessed at https://scminer.stjude.org/home.
Bottom Line
Like the gangster film crime bosses whose involvement in a crime may only be discoverable through the actions of their associates, some cancer drivers are not detectable using direct profiling methods alone. scMINER ferrets out these hidden drivers using network-based approaches that implicate genes based on the expression of their downstream targets. Whether you’re a researcher or simply curious about how cutting-edge tools are transforming science, scMINER‘s innovative approach to discerning the molecular networks underlying disease is a useful addition to the cancer research toolbelt.