Wave Life Sciences and Deep Genomics have joined forces to find therapies for genetic neuromuscular disorders. Deep Genomics’ artificial intelligence platform can predict the effects of patient variants on RNA processing, as well as create steric blocking oligonucleotides to modify splicing.
This system helped identify the mechanism underlying Met645Arg, which led to exon skipping and ATP7B dysfunction in Wilson disease. Discovery took only 18 months compared with traditional biopharma drug development which can take between 3-6 years.
Using AI to accelerate drug discovery
Drug discovery poses many difficulties in terms of understanding how to manipulate genes and proteins. To overcome this hurdle, researchers are turning to AI technology as a solution – specifically artificial neural network technology which enables researchers to predict how a new chemical will interact with target proteins – in order to accelerate drug discovery processes and decrease testing requirements while increasing success rates of drug development.
Studies have demonstrated the potential benefits of AI drug target identification technology to speed up drug discovery. AI technologies also tend to produce more accurate predictive models than conventional methods; however, their accuracy may be limited by available experimental data for training them – companies developing such systems must ensure sufficient experimental data exists so as to improve predictions that will ensure safety and efficacy for patients.
Deep Genomics offers an artificial intelligence platform designed to support geneticists, molecular biologists, chemists, toxicologists, drug developers and more. Deep Genomics uses machine learning techniques to detect hard-to-detect disease triggers and develop genetic medicines. Their team is comprised of PhDs with advanced degrees in fields ranging from AI/automation through cell and molecular biology/in vitro disease modeling/clinical development – located within Toronto’s MaRS Discovery District near four research hospitals/medical institutes as well as Google/Uber/Vector Institute of AI labs/Institute.
AI Workbench drug discovery platform from the company helps identify potential oligonucleotide therapies that target genetic determinants of disease at both RNA or DNA levels, along with any off-target effects, cell viability data or animal toxicology data generated. Furthermore, this platform generates on-target and genome-wide off-target effects data as well as cell viability results; its AI model learns from these results to improve performance and accuracy over time.
Deep Genomics recently joined forces with Wave Life Sciences to identify novel treatments for hereditary neuromuscular disorders. Their collaboration will speed the discovery and development of antisense oligonucleotide therapies targeting exon 51 for DMD, such as WVE-210201 from Wave’s WVE-210201 program.
AI’s use in healthcare to identify rare diseases is an exciting development for pharmaceutical industry and could provide breakthroughs that ultimately improve lives of millions. Yet implementing AI in this space presents several obstacles; most importantly ensuring privacy and security are respected while using it wisely in collaboration with expert oversight and collaboration.
Developing oligonucleotides for neuromuscular disorders
At present, numerous antisense oligonucleotide (ASO) therapies are currently under development for genetic neuromuscular disorders. These drugs aim to target mutations and their associated cellular processes in order to alter them for positive results; such ASOs could even restore normal muscle function for some patients with hereditary diseases by decreasing symptoms or improving motor function.
ASOs work by inhibiting the splicing of pre-mRNA transcripts, which are then degraded and excreted from the body – this mechanism can be leveraged to treat various diseases including DMD and SMA, while also improving quality of life while slowing disease progression.
Deep Genomics team leader Brendan Frey recently published a preprint in bioRxiv detailing an approach for identifying optimal genes to target with stereopure oligonucleotides using machine learning systems such as AI Workbench as well as large amounts of proprietary RNA data. Splicing of RNA plays an essential role in neuromuscular disorders caused by genetic mutations.
This technology enables more effective drug discovery and helps scientists better understand how to treat rare neuromuscular disorders. However, machine learning findings are only as accurate as the data used. Therefore, it is crucial that genomic data derived from public sources or research populations that reflect wider society be included for accurate machine learning analysis.
Achievement of optimal therapeutic efficacy with minimal toxicity remains one of the greatest obstacles to accelerating clinical development, necessitating greater knowledge about genetic underpinnings of disease and prioritizing oligonucleotides for optimal targets in order to ensure these therapies reach clinic.
Recent advancements in oligonucleotide therapy include the creation of an ASO-based treatment for Wilson disease, an inherited disorder in which copper accumulates in liver, brain and other organs. A study using the DG12P1 oligonucleotide revealed it successfully reduced copper accumulation among those with Met645Arg Wilson disease variant, where function loss of the ATP7B copper-binding protein occurs; findings consistent with previous preclinical research using this compound with Wilson disease mice and human hepatoma cells.
Developing oligonucleotides for cancer
Oligonucleotides provide an alternative to conventional cancer chemotherapy drugs by targeting specific genes or regions instead of all cells simultaneously, making them far less toxic and rapidly eliminated from the body. Unfortunately, however, creating effective oncological oligonucleotides requires considerable precision and affinity in their design in order to effectively target tumor cells without entering bloodstream and creating systemic effects.
Scientists must fully comprehend how gene-produced protein functions for effective oligonucleotide therapy to be developed safely and successfully. This protein, known as a transcription factor, regulates gene expression by binding to and blocking or activating certain genes. Cancer cells produce abnormal amounts of transcription factors due to mutations that alter how this gene produces its protein; as a result, cancerous cells grow unchecked while also multiplying uncontrollably.
Oligonucleotides have become essential components of cancer treatments. These small molecules mimicking RNA can bind with high specificity and affinity to their target molecules, inhibiting protein activity by targeting their messenger ribonucleic acid (miR), or used antisense approach to suppress or even eliminate gene expression.
Deep Genomics of Toronto recently secured $180 Million through its Series C round, more than quadrupling their previous $40 Million fundraising. With this sum in hand, at least 10 computer-designed drugs will enter clinical trials, using Deep Genomics’ AI platform which uses short strips of genetic material as targets in targeting disease RNA or DNA molecules.
Deep Genomics will use this new funding to expand their AI Workbench program, which helps biopharma companies transform RNA into potential therapeutic compounds. Their AI model uses proprietary data accumulated over decades to address the complex biology behind RNA’s structure and identify targets, mechanisms and therapeutic compounds for potential development into potential treatments.
Wave and Deep Genomics’ collaboration aims to identify new classes of anti-sense oligonucleotide therapies for metabolic and neurodegenerative diseases. Both companies will leverage their respective expertise in splice correction to produce stereopure oligonucleotides targeted to specific regions of a genome for efficient therapeutic effects with reduced off-target effects.
Developing oligonucleotides for rare diseases
Therapeutic oligonucleotides work at the genetic level to interfere with gene expression, making them powerful weapons against rare diseases. They come in many forms such as antisense oligonucleotides (ASO), small interfering RNA (siRNA), and microRNA, all designed to either disrupt or modulate gene expression and correcting the misproduction of proteins that leads to genetic diseases.
Development of oligonucleotides to treat rare diseases can be particularly challenging as these drugs must often be personalized for a limited number of patients. Such personalized medication requires extensive clinical knowledge about both affected individuals and symptoms as well as thorough testing to ensure an oligonucleotide is effective against its target condition.
Pharmaceutical companies play an essential role here. An estimated 3.5 million people in the UK suffer from rare diseases, with 95% having no viable treatment options available to them. Patients need highly tailored therapies developed quickly that target each individual; targeting large groups would not be feasible or efficient enough. A promising approach to meeting this challenge are n = 1 antisense oligonucleotides such as Milasen for treating Batten disease is now in production – its first use being with seven year-old girl suffering from this severe metabolic condition was produced and prescribed to her in 2018.
Still, developing ASOs for individual patients remains challenging. Genomics-wide assays and high-throughput molecular biology data must be utilized in order to select appropriate targets and regions within those targets for intervention using an oligonucleotide intervention oligonucleotides must then be made stereopure before being delivered with specific delivery technologies, which is an expensive endeavor.
Deep Genomics faces an enormous challenge in maintaining its machine learning thought leadership while pursuing drug development and pharmaceutical partnerships. To do so successfully requires careful resource allocation to maintain their core platform and robust intellectual property portfolio. Furthermore, ongoing public data leveraging is vital in order to secure short-term pharma partnerships while building long-term competitive advantage for their company.
