New machine-learning tech 'reads and interprets data like a PhD scientist'
Canada-based BenchSci, recently announced the launch of ASCEND that it says is the first map of all disease biology aiming to transform pharmaceutical research. It says it is powered by machine learning technology that reads and interprets data like a PhD scientist.
According to the company, it will remove the barriers that result in 98% of pharmaceutical research investment that fails to reach patients.
BenchSci believes this will move the most promising projects forward fast and says the technology was trained by scientists to extract experimental evidence from internal and external sources. It uses curated ontology datasets and makes connections across experiment outcomes to create the first commercially available, unbiased and evidence-based map of the underlying biology of disease.
Philip Tagari, vice president of research at Amgen and his team worked closely with BenchSci to shape the technology over two years. Today, all therapeutic area preclinical teams at Amgen can leverage this novel platform to advance their work.
Transform speed and success
“BenchSci has developed a technology with the potential to transform the speed and success of preclinical research,” Tagari said .
“With ASCEND, BenchSci has developed a unique approach to extracting and connecting scientific evidence from within and outside Amgen. Based on results across many programs, we are excited to continue to provide ASCEND broadly across the organization to catalyze our research.”
BenchSci says that top pharmaceutical companies that were early adopters of ASCEND found significant improvements to their portfolio performance which included 40% of projects identified a new indication to explore or an additional target gene not previously considered.
Research and development productivity was improved by 33% in projects identified with a safety or efficacy risk.
Dramatic reduction in unecessary experiments
Retrospective analysis unveiled that scientists could have uncovered insights and dramatically reduced unnecessary experiments to accelerate pre-clinical programs by a minimum of 40% improvement in this phase of drug development
ASCEND, an end-to-end enterprise-wide software as a service (SaaS) solution, guides scientists at every stage of preclinical research by augmenting target selection, due diligence and hypothesis generation, BenchSci says.
It also augments target selection, due diligence and hypothesis generation and develops optimal investigative approaches to test hypotheses and design experiments that yield definitive results and reduce trial-and-error.
The company says it can also identify safety and efficacy risks to support successful IND (investigational new drug) submission and clinical translation
Through the eyes and mind of scientists
Liran Belenzon, CEO and co-founder of BenchSci said: “We share our partners’ visions to help bring hope to patients faster. Our role in solving this enormous challenge is to develop and train technology that can change the world through the eyes and mind of scientists,”
“It’s not simply the proprietary AI that’s revolutionary. What’s remarkable about ASCEND is the unification of cutting-edge technology, a depth of experience in disease biology and our collaboration with leading pharmaceutical companies that has created the potential to advance the speed and success of better medicine to patients.”
In its research BenchSci found that on average, 98 percent of pharmaceutical experimental research projects, including the respective time, materials, and brainpower deployed in the quest for novel medicines, fail to reach patients. This issue costs the industry billions in research and development investment annually and delays new medicines for patients in need.
BenchSci said he increasing complexity of disease biology makes finding novel discoveries challenging. Scientists have been underserved for decades without major advancements in tools and technology to efficiently navigate the magnitude of scientific data and evidence.