Deep Genomics of Toronto believes artificial intelligence should do much of the work that humans currently perform when developing new medicines to treat genetic diseases. They have formed a partnership with Wave Life Sciences from Cambridge and plan to use their expertise in hereditary neuromuscular disorders as their focus area.<\/p>\n
Wave has agreed to use Deep Genomics’ AI Workbench, which integrates automation, data science, biomedical knowledge and machine learning technologies. According to Deep Genomics’ claims about their system’s efficacy in targeting therapeutic targets previously regarded as undruggable; furthermore it has proven itself effective at shortening drug discovery timelines and increasing clinical trial success rates.<\/p>\n
Deep Genomics’ AI Workbench is an advanced suite of systems that helps researchers select suitable research candidates, assess potential therapeutic mechanisms, design molecules to assess potency and safety, design animal models, assess patient subpopulations and select an ideal treatment regimen for each subpopulation. While other AI companies apply their algorithms only to certain parts of pharmaceutical processes, Deep Genomics has built its business around incorporating AI as part of an overall workflow solution.<\/p>\n
Deep Genomics claims its technology can be used to rapidly identify novel RNA therapeutics using machine learning-powered approaches. RNA therapies are nucleic acid-based medicines which are delivered into cells to modify their function; two have already been approved and many others are under development.<\/p>\n
RNA therapeutics can be developed through various approaches, including gene editing, small molecule drugs, cell therapies and ribozymes. Unfortunately, there can be several significant challenges involved with taking these therapies to market – one being its efficacy for humans, which depends on numerous factors. Deep Genomics’ AI Workbench uses machine learning technology to predict which therapies may be effective against potential patient populations.<\/p>\n
Deep Genomics was established in 2015, employing 40 employees with degrees and expertise in artificial intelligence, automation, cell and molecular biology, medicine, preclinical development, drug discovery organic chemistry software engineering. Their team has worked for years on an enterprise platform designed to be fully integrated with their operations. Deep Genomics held talks with many major pharmaceutical companies before choosing Wave Biotech because its research-oriented startup mentality resonated with Deep Genomics’ goals.<\/p>\n
Recent research has demonstrated that pharmaceutical industry researchers spend on average $2.6 billion dollars and 10 years conducting research and development activities to produce one approved drug, largely due to trial and error processes in traditional R&D activities. With sequencing costs decreasing and large genomic datasets becoming readily available, artificial intelligence holds immense promise as an efficient means of cutting both time and expenses associated with traditional R&D activities.<\/p>\n
Deep Genomics, established in 2015, specializes in using machine learning to help scientists discover druggable mutations more quickly and accurately. Their proprietary discovery technology integrates automation, biomedical knowledge, and machine learning expertise for therapeutic targets that might otherwise remain undruggable.<\/p>\n
The platform utilizes both proprietary and publicly available genetic sequence data to gain insight into gene function and search for disease-causing mutations, prioritizing those for further study as possible drug targets. AI Workbench software from the company helps researchers rapidly evaluate compounds designed to target those mutations while automatically creating an oligonucleotide design list with potential therapeutic impact.<\/p>\n
Deep Genomics is currently capitalizing on its machine learning thought leadership to secure pharmaceutical partnerships that will assist its drug development efforts. Therefore, the company should remain wary not to divert resources away from its core machine learning platform and proprietary data which have helped establish its presence in the market. Furthermore, data collection must be representative and valid; machine learning results rely heavily on quality underlying data sources; therefore Deep Genomics should look for partnerships with pharmaceutical firms who provide access to quality genomic and clinical data that is continually being updated and refined.<\/p>\n
Human beings cannot make sense of all the biological data generated by experiments without bioinformatics, an emerging field combining domain science knowledge, mathematical formulae and statistics with computer software for data interpretation and interpretation. Bioinformatics seeks to find patterns in datasets while creating mathematical models of how these parts of biological systems such as genes, proteins, bacteria, plants animals or ecosystems interact.<\/p>\n
Bioinformatics began as a means to decipher molecular sequences that comprise proteins and understand their function in cell processes, helping researchers identify genes linked to traits seen in humans (phenotypes). Now bioinformatics can scour large databases looking for patterns in biological data which could explain genetic diseases or indicate whether an infectious illness will respond well to antibiotic treatments.<\/p>\n
Scientists can now sequence much more DNA, thanks to computational tools developed for this process and advances in computer power, making bioinformatics more accessible than ever before. But bioinformaticians must remain mindful of its limitations: machine learning algorithms only learn as accurately as the quality of data from which they were trained on; hence it is imperative that genomic data sets come from high-quality and representative research populations.<\/p>\n
PNNL’s bioinformatics researchers are creating tools to more effectively capture, organize and utilize this data. Their efforts are also used to help develop an improved forensic tool using protein profiles instead of DNA for use in cases when suspects’ DNA may be absent or unclear; additionally forensic proteomics could potentially assist law enforcement officials with detecting drugs, toxin-based toxins and sports doping hormones even when their DNA has been deleted from a body or destroyed.<\/p>\n
Bioinformatics has become an indispensable part of everyday life as its popularity and impact grow exponentially. Bioinformatics helps scientists better understand the genetics behind diseases, predict phenotypes from genome sequencing data and identify new drug targets to deliver more personalized medicine treatments. Furthermore, it also aids us in better comprehending cell metabolism; fat, protein and metabolite transport within cells as well as assess human microbiomes.<\/p>\n
Deep Genomics’ automated platform leverages its own patented machine learning systems and processes, taking as input a test variant’s DNA or RNA sequence, extracting features, comparing them against known disease-causing variants, and selecting candidate lists for further evaluation. Their platform continually evolves by adding data from rapidly expanding public sources that enhance accuracy; on-target and off-target effects from its system are fed back in as feedback for further tuning.<\/p>\n
Deep Genomics, established in Toronto in 2015, specializes in finding hard-to-detect disease triggers and developing drugs to treat hereditary conditions. Their team boasts extensive industry experience as well as advanced degrees in areas like artificial intelligence, automation, medicine, cell and molecular biology, in vivo disease models, drug discovery preclinical development organic chemistry software engineering.<\/p>\n
As evidenced in a preprint published by bioRxiv, Deep Genomics’ AI Workbench successfully identified and prioritized NM_000053.3:c.1934T>G – an ultra rare Wilson disease variant with a loss-of-function mutation in ATP7B copper-binding protein – among patients, helping prevent life-threatening accumulation of copper ions that could harm livers and brains. Deep Genomics offers their oligonucleotide therapy called DG12P1 designed specifically to correct this mutation in patients, helping these individuals avoid life-threatening accumulation of copper ions which could damage organs as well.<\/p>\n
Deep Genomics’ R&D costs continue to skyrocket, so companies such as Deep Genomics are seeking pharmaceutical partnerships in order to expedite drug discovery processes more rapidly. Yet finding these alliances may prove challenging as recent studies indicate the average pharmaceutical company spends between $2.6 billion and 10 years developing one new drug; much of this cost stems from traditional trial-and-error processes associated with traditional drug discovery.<\/p>\n
Companies seeking to cut costs and accelerate drug discovery processes are increasingly turning to artificial intelligence (AI) techniques in their pipelines, specifically machine learning techniques such as Deep Neural Nets to help identify promising candidates, narrow down potential drug targets for further study, predict biological outcomes for hereditary diseases and identify promising biological targets for further investigation. Thanks to decreasing sequencing costs and more genomic data sets being made available for analysis, these AI-enabled discovery platforms are dramatically decreasing time and costs when developing new medicines.<\/p>\n