fbpx How Precise is Your Medicine? - Blog Post | JLABS

How Precise is Your Medicine?


In the decades since the human genome was first decoded, dozens, if not hundreds of targeted medicines have become available. What’s increasingly become clear – as we shift from population medicine to precision medicine – is the need to better understand what combinations of these new therapies can help prevent, manage or cure disease.

What’s the best way to sort through all the existing studies, data and genomics to determine the best approach for patients? And, as we’re learning that combinations of these biomarker-specific medicines can sometimes be more effective than single drugs, how can doctors determine the best mix for each patient?

Several startups at Johnson & Johnson Innovation, JLABS are working on that specific area. We caught up with three of these companies, residing at JLABS @ San Diego, ahead of the Precision Medicine Leadership Summit, taking place the week of August 21st.


If you’ve ever used a travel app that coordinates your plans from different hotels, airlines and rental car companies into a single itinerary, then you have a basic understanding of what CureMatch is trying to do for oncology.

The digital health company’s system aims to help doctors select the best combinations of cancer drugs for individual patients, based on the molecular profile of the person’s own tumor. With 650 known cancer-related genes and 300 FDA-approved medicines to treat the disease, finding the right mix of therapies can be daunting. Even the top oncologists would have difficulty assessing the possible 4.5 million three-drug combinations that would work best for each patient, based on that person’s genetic data.

CureMatch’s software analyzes the genetic sequencing information from a patient’s tumor, compares it with a database of other patients, clinical trials, published studies and expert medical advice, and scores the combinations of drugs most likely to work. The top five results are then detailed in a report sent to the patient’s oncologist for review.

“Combining different therapies has been successfully used to treat diseases such as HIV/AIDS, but unfortunately, a similar approach isn’t as widely available for cancer patients,” said Stephane Richard, president and COO of CureMatch. “We make it possible for oncologists to see what combination of targeted therapies, hormone therapies, immunotherapies and other cancer drugs would likely be the most effective for each of their patients.”

The service is available to all patients–adults and children–with solid or hematologic tumors.


For some cancer patients, immunotherapy has been a literal life-saver. But these medicines only work for a fraction of people afflicted with the disease.

Salgomed, which is named after the phrase “Search Algorithms for Medicine,” aims to use artificial intelligence to point doctors and pharma companies toward future medicines and drug combinations that will boost the number of patients who can benefit from these therapies.

“We asked, ‘what can we do for patients that don’t respond to immunotherapy? Can we create better drugs and give them to the right patients?’” said Salgomed CEO Giovanni Paternostro.

To find a better solution, Salgomed starts by building computational network models of immune and cancer cells from genomic data. These models indicate what drug combinations would likely have the best impact. Then, using its algorithm-driven robotics, Salgomed tests thousands of these combinations in an experimental platform based off a patient’s cells. The results can be integrated with public datasets and used to refine the network models, leading to constantly improving cycles of computation and experimentation.

Ultimately, the goal is to identify novel cancer immunotherapy drug targets and to provide rapid clinical validation of candidate drugs.

“If you really look at how immune system works, it has the potential to be used to fight a large fraction of human diseases,” Paternostro said. “Combinations of new and existing immunotherapies will play a key role in the future of medicine.”

DNA-SEQ Alliance

One of medicine’s biggest issues in the genomic age is deriving meaningful insights from the growing flood of emerging genomic data. The number of biomarkers, mutations and potential targets linked to disease is exponentially increasing. Some may be irrelevant, others may only be useful for certain patients, and still others may only be effective in combination – but all of these insights can now inform the treatment and discovery process.

DNA-SEQ Alliance was formed to better understand this flow of genomic data. The group is taking advantage of advances in rapid genome sequencing, precise genome analysis and mapping mutations to aberrations in molecular structures. The goal is to hone in on better personalized medical recommendations for patients and to produce precise templates that will improve drug design.

By using disruptive artificial intelligence capabilities such as pattern matching and machine learning to augment the precision of crystallography, the Alliance has designed an analytic engine that will ultimately be transferred to one of the world’s first quantum computers built by D Wave Systems. DNA-SEQ aims to utilize these novel capabilities to improve cancer diagnostics and drug discovery.

“Based on that approach we have made two fundamental discoveries on how cancer actually hijacks the mechanism related to cell signaling,” said Janusz Sowadski, Founder & CEO of DNA-SEQ Alliance. “Now we are using these findings to design a novel class of medicines.”

Sowadski, who was trained as a crystallographer – a scientist focused on the crystal structures of proteins - “solved” the first structure of protein kinases in 1991 with his team in La Jolla, California. That discovery established a firm structural foundation for the entire extended family of specific cancer cell inhibitors that exists today.

DNA-SEQ has a library of more than 2,000 structures containing more than 7 million XYZ coordinates. This kinase crystal library is the organizing principle of the Alliance pipeline’s data gathering cycles. The algorithm’s pattern matching capabilities, coupled with the “teaching lessons” that have been embedded through machine learning, can then more precisely map the mutation by searching the 3D crystal space, and match it to an optimal therapy.

The alliance’s goal is to use this system to provide oncologists and pharma companies with a technological support system to practice a more precise brand of personalized medicine, helping to transform cancer into a manageable disease.