Imagine turning what doctors often dismiss as a pesky side effect into a lifesaving treatment for another patient—that’s the groundbreaking promise of a new tool called DeepTarget, revolutionizing how we fight cancer.
A fresh research paper published in Nature (https://www.nature.com/articles/s41698-025-01111-4) by experts at Sanford Burnham Prebys Medical Discovery Institute (https://sbpdiscovery.org/) reveals an exciting twist: side effects that trouble one person might actually hold the key to healing someone else. The secret? Widening our view of how small molecule drugs—those tiny chemical compounds that make up a huge chunk of our medications—interact with the body. For beginners, think of small molecules as the building blocks of many pills we take; they’re not like natural proteins but synthetic helpers designed in labs to tweak our biology in precise ways.
Small molecules: versatile players in the body
When we zoom out and consider the full landscape, it’s clear that these small molecules don’t stick to just one job. Depending on the illness or the type of cells involved, they can latch onto various proteins and trigger different outcomes. This flexibility opens doors for reusing existing drugs in fresh ways, potentially helping way more people get effective therapies. And this is the part most people miss: instead of seeing these extra interactions as mistakes, we could harness them as hidden strengths.
“Many of the small molecules in our everyday medicines aren’t something you’d find naturally occurring in the wild, so they didn’t evolve for a single, narrow purpose,” explains Dr. Sanju Sinha, an assistant professor specializing in cancer metabolism and the tumor microenvironment at Sanford Burnham Prebys. “Too often, scientists zero in on these drugs as if they only hit one main target, with any other effects brushed off as unwanted ‘off-target’ glitches.” But here’s where it gets controversial: is labeling them as glitches holding back innovation, or is that caution what keeps treatments safe? It’s a debate worth pondering.
Enter DeepTarget: revolutionizing target prediction
Dr. Sinha’s journey into the adaptable nature of small molecule drugs started during his time at the National Cancer Institute. There, he crafted DeepTarget, a smart computational tool that doesn’t rely on a drug’s chemical blueprint to guess its targets. Instead, it dives into massive troves of genetic experiments and drug response data from cancer cells. To give you a sense of scale, the analysis covered 1,450 different drugs tested on 371 types of cancer cell lines, all pulled from the reliable Dependency Map (DepMap) Consortium—a collaborative effort that maps out what genes cancer cells depend on to survive.
Why does spotting these backup targets matter so much? Well, a ton of drugs already approved by the FDA, plus many in the pipeline for clinical trials, come with these secondary effects baked in. For newcomers to this field, primary targets are the main proteins a drug is designed to block or activate, while secondary ones are the unexpected bonuses (or surprises) it hits along the way.
In head-to-head tests, DeepTarget shone brighter than top existing tools like RoseTTAFold All-Atom and Chai-1, coming out on top in seven out of eight key challenges. It nailed predictions for both standard and mutated versions of target proteins in cancer, and even pinpointed those elusive secondary targets with impressive accuracy. “Predicting these extra targets is a game-changer since so many approved and developing drugs have them,” notes Sinha, the study’s lead author. “Rather than treating them like flaws, if we view them as useful traits, we unlock huge potential for repurposing drugs to fit new needs.”
Real-world proof: Ibrutinib’s surprising role in lung cancer
To put DeepTarget through its paces, the team ran hands-on experiments, including a compelling look at Ibrutinib. This drug is already a go-to FDA-approved option for certain blood cancers, where it primarily blocks Bruton’s tyrosine kinase (BTK), an enzyme that helps cancer cells grow. But earlier patient studies hinted it might also work against lung cancer—even though BTK isn’t a factor in those solid tumors. How could that be?
Teaming up with Dr. Ani Deshpande, a co-senior author and professor in the Cancer Genome and Epigenetics Program at Sanford Burnham Prebys, the researchers used DeepTarget to probe if Ibrutinib was secretly targeting epidermal growth factor receptor (EGFR) in lung cells. EGFR is a protein often mutated in lung cancers, driving uncontrolled cell growth like a faulty accelerator in a car. For clarity, mutations here mean changes in the DNA that make the protein hyperactive, fueling the disease.
“When we fed DeepTarget data focused on blood cancers, BTK popped up as the star target,” Sinha shares. “But switch the lens to solid tumors like lung cancer, and suddenly a mutated, cancer-promoting version of EGFR takes center stage. It’s a perfect illustration of how targets can shift based on the context—like the same tool working differently in various workshops.” Sure enough, lab tests showed lung cancer cells with the faulty EGFR were extra vulnerable to Ibrutinib, solidifying EGFR as a valid secondary bullseye.
At its core, DeepTarget’s smarts come from a clever biological mimicry: it assumes that knocking out a gene (using CRISPR-Cas9 editing, a precise ‘cut-and-paste’ technique for DNA) that codes for a drug’s target protein should produce effects similar to the drug itself. This approach draws from those expansive datasets of 1,450 drugs on 371 cell lines, giving it a robust foundation. (Image credit: Sanju Sinha, Sanford Burnham Prebys.)
Broader impacts on crafting better drugs
“What makes DeepTarget stand out in practical settings is how it echoes the true complexities of drugs in action,” Sinha adds. “Real biology isn’t just about a molecule snapping onto a protein; it’s about the whole cellular environment and interconnected pathways influencing the results.” This holistic view could speed up creating and adapting drugs, working hand-in-hand with traditional methods that obsess over exact molecular fits.
The research signals a big mindset shift in drug discovery: moving beyond rigid ‘one drug, one target’ thinking to embrace the nuances of cell types and disease stages. By doing so, innovations like DeepTarget might fast-track novel therapies. Looking ahead, Sinha is eager to apply these findings to design entirely new small molecule treatments.
“Boosting options for cancer care—and tackling tougher challenges like aging—will hinge on refining our grasp of life’s inner workings and fine-tuning therapies to match,” he concludes.
But let’s stir the pot a bit: while this context-specific approach sounds revolutionary, could over-relying on secondary targets risk unintended consequences in patients? Or is the real controversy in how slowly the pharma world adopts these ideas? What do you think—should we embrace the ‘bugs as features’ philosophy, or stick to safer, single-target strategies? Drop your thoughts in the comments; I’d love to hear if you’re team repurposing or team caution!