1. ASO (Antisense Oligonucleotide)
A short synthetic strand of DNA designed to specifically bind to target mRNA and block its function, leading to cancer cell death.
We aim to checkmate lung cancer with a modular therapy that combines targeted delivery (antibody), precise gene silencing (ASO), synthetic-lethality logic, and immune activation (epitope presentation), all guided by a computational design engine. Short-term, we deliver in-vitro validation; long-term, we build toward a first-in-human path for NSCLC.
Cancer remains one of the leading causes of death worldwide, with lung cancer ranking among the most aggressive and deadly types. According to the American Cancer Society (2025), lung cancer accounts for roughly 1 in 5 cancer-related deaths in the United States, with more than 220,000 new cases and over 120,000 deaths expected this year [1]. Globally, GLOBOCAN 2022 reports nearly 2.5 million new lung cancer cases (12.4% of all cancers) and almost 1 in 5 cancer deaths [2]. Despite advances in detection and therapy, non-small cell lung cancer (NSCLC), which represents ~85% of cases, is still frequently diagnosed at late stages and has poor long-term outcomes [3].
Beyond the numbers, lung cancer profoundly impacts patients’ daily lives. Symptoms such as persistent cough, chest pain, shortness of breath, and fatigue often progress rapidly, limiting physical activity and independence [4]. The 5-year survival rate for stage IV lung cancer is only about 10%, underscoring its dire prognosis. In advanced stages, patients often experience significant weight loss, pain, and frequent hospitalizations, leading to a steep deterioration in quality of life [5]. Emotional distress, anxiety, and depression are common, both for patients and their families, as treatment courses are long, complex, and rarely curative [5],[6]. Even when therapies extend survival, they are often accompanied by severe side effects, creating a dual burden of living with both the disease and the toxicity of its treatment. This stark reality underscores the urgent need for therapies that are not only more effective but also gentler and more sustainable for patients [7].
These newer therapies have improved outcomes, but their benefits remain limited. Most drugs target only cell-surface proteins, leaving many intracellular drivers inaccessible. Resistance mutations frequently emerge, making once-effective therapies obsolete within months. Eligibility is narrow, since treatments often apply only to patients with very specific mutations. And perhaps most importantly, loss-of-function tumor suppressors such as TP53 or RB1, among the most common alterations in lung cancer, remain essentially untreatable with current approaches [8],[9],[10],[11],[12].
To read more about the Problem and Unmet need -> Click Here.
Cancer is not a single, uniform enemy - it evolves, hides, and adapts. Tackling it requires more than one weapon; it demands a combination of precision tools and diverse strategies. Much like in chess, success comes from attacking from multiple directions until the opponent is cornered - our mission is to Checkmate Cancer.
We introduce our five main “chess pieces,” each representing a different aspect of our approach:
1. ASO (Antisense Oligonucleotide)
A short synthetic strand of DNA designed to specifically bind to target mRNA and block its function, leading to cancer cell death.
Antisense oligonucleotides (ASOs) are short, synthetic single-stranded nucleic acids, usually between 15 and 25 nucleotides in length, that are designed to bind complementary messenger RNA (mRNA) sequences inside human cells [14]. They work by taking advantage of the universal rules of Watson–Crick base pairing, which allow them to selectively recognize one RNA target among thousands. Once bound to their chosen RNA, ASOs can interfere in several ways: they may block its use, alter how it is processed, or trigger its degradation, ultimately leading to a reduction or complete loss of the corresponding protein [15].
What makes ASOs different from conventional drugs is the stage at which they act. Most medicines, such as small molecules or antibodies, target proteins directly blocking or neutralizing them after they are already produced. ASOs act one step earlier, at the RNA stage, preventing the protein from ever being made. This upstream mode of action enables ASOs to silence disease-causing genes before their protein products appear, opening the door to therapeutic strategies that were previously considered impossible for traditional drugs [15].
The idea of antisense therapy was first demonstrated in 1978, when Paul Zamecnik and Mary Stephenson showed that a 13-nucleotide DNA fragment could specifically block viral protein production by binding to RNA from Rous sarcoma virus [14]. Twenty years later, the first FDA-approved antisense medicine, fomivirsen (Vitravene), was introduced for the treatment of cytomegalovirus retinitis. Since then, antisense therapy has moved from concept to clinical reality, with several additional ASO drugs approved: nusinersen (Spinraza) for spinal muscular atrophy [16], eteplirsen (Exondys 51) for Duchenne muscular dystrophy [17], inotersen (Tegsedi) for hereditary amyloidosis [15], and volanesorsen (Waylivra) for familial chylomicronemia syndrome [15]. These milestones firmly establish ASOs as a validated therapeutic platform with broad potential in modern medicine.
ASOs can influence gene expression through several distinct mechanisms, each of which changes the fate of the target RNA in a different way.
The best-known mechanism is RNase H-mediated degradation. When an ASO binds to its target mRNA, the two strands form a DNA:RNA hybrid. This hybrid is recognized by RNase H1, an enzyme presented in both the nucleus and the cytoplasm, which selectively cleaves the RNA strand while leaving the ASO intact. Because the ASO survives the process, it can go on to bind and destroy additional RNA molecules, working in a catalytic fashion [18]. Over time, this results in a substantial reduction of the target RNA and its encoded protein. Importantly, not all binding sites are equally effective. The sequence context surrounding the ASO binding site strongly influences how well RNase H1 can act, with certain motifs promoting cleavage and others reducing it. Understanding these preferences is a central part of rational ASO design.
Another powerful mechanism is splicing modulation. Before mRNA is ready to be translated, it is first transcribed as pre-mRNA, which contains both exons (coding regions) and introns (non-coding regions). This transcript must be processed by the spliceosome, which cuts out introns and joins exons together. By binding to specific sites on the pre-mRNA, ASOs can block the spliceosome and force the cell to include or skip selected exons. This redirection of splicing can change the final protein product. Clinically, this approach is exemplified by nusinersen (Spinraza), which corrects defective splicing of the SMN2 gene in spinal muscular atrophy, thereby restoring the production of functional SMN protein [16].
ASOs can also act through steric blocking, a mechanism that is conceptually simpler but often just as effective. In this case, the ASO binds to an mRNA at a site that is crucial for translation or for the binding of regulatory proteins. Its physical presence alone is enough to block access, preventing the ribosome from initiating or continuing protein synthesis. Unlike RNase H-mediated degradation, the RNA itself remains intact, but it can no longer be used to produce protein [15].
Together, these mechanisms give ASOs remarkable versatility: they can destroy an RNA entirely, correct its processing, or simply block it from being used. This diversity of action is one of the reasons ASOs have become such a powerful therapeutic platform.
Designing an antisense oligonucleotide is not as simple as writing down the reverse complement of a target RNA. Messenger RNAs are highly structured molecules that fold into loops, stems, and complex three-dimensional shapes. Some regions are hidden within double-stranded stems, while others remain single-stranded and exposed. For an ASO to work efficiently, it must bind to one of these accessible regions where pairing can take place without strong structural barriers [19].
The strength of the ASO–RNA interaction must also be carefully tuned. If binding is too weak, the ASO will fall off before it has an effect. But if binding is too strong, the duplex can become overly rigid. This may reduce the ability of RNase H1 to recognize and cut the RNA or increase the chance that the ASO will stick to unintended transcripts that share partial sequence similarity [15]. Finding this “sweet spot” in binding strength is part of the art of ASO design.
Another factor is the local sequence context. RNase H1, the enzyme that drives RNA cleavage, does not cut every site equally well. Certain nucleotide patterns, both single bases and dinucleotides near the cleavage point, strongly influence its efficiency. Targeting regions that favor RNase H1 recruitment can make the difference between an ASO that works in theory and one that is effective in practice [18].
Specificity is equally critical. The human transcriptome is vast, and many genes share short stretches of similar sequence. Without careful selection, an ASO may accidentally bind to unintended RNAs, leading to off-target effects and possible toxicity. Computational screening is therefore a critical part of any rational ASO design [19].
In our project, these biological considerations are not treated as abstract knowledge. Instead, they are transformed into quantitative features that our computational pipeline can measure and compare. RNA structure, hybridization strength, sequence motifs, and specificity are systematically analyzed, and together they inform the selection of ASOs with the highest likelihood of success. The technical details of how we implement these features are presented in the modeling section, where we describe our predictive algorithms in depth. Read more about our Model.
When designing ASOs, chemical modifications play a central role. Without them, oligonucleotides are inherently unstable: in biological fluids they are quickly degraded by nucleases, giving them a very short half-life and poor bioavailability. Chemical modifications solve this problem. By altering the backbone, sugar, or nucleobase, ASOs can remain stable in biological environments, reach their target sites, and maintain therapeutic activity [20]. At the same time, the type and distribution of these modifications strongly influence potency, specificity, and safety. In other words, chemical design is not optional-it defines the therapeutic window of ASOs.
One of the earliest and most widely adopted modifications is the phosphorothioate (PS) backbone. In this chemistry, one of the non-bridging oxygen atoms in the phosphate group is replaced by sulfur. This subtle change dramatically improves nuclease resistance, helping ASOs persist in biological fluids. PS linkages also promote interactions with plasma and cellular proteins, improving systemic distribution and uptake into cells. However, these benefits come with trade-offs: PS modifications can slightly reduce RNA binding affinity and are associated with immune-related toxicities, such as complement activation or platelet effects [21],[22]. Thus, while PS chemistry forms the backbone of almost all therapeutic ASOs, it illustrates the balance between improved pharmacokinetics and potential safety liabilities.
Sugar modifications form another major class of strategies. Substitutions at the 2′ position of the ribose, such as 2′-O-methyl (2′-OMe) and 2′-O-methoxyethyl (2′-MOE), further enhance nuclease resistance and improve RNA binding affinity. These modifications also reduce immune recognition, which is critical for tolerability in vivo. Among sugar chemistries, Locked Nucleic Acids (LNAs) are particularly notable. LNAs include a methylene bridge that locks the ribose into a rigid conformation, dramatically increasing the binding affinity for RNA and the thermal stability of the ASO-RNA duplex [21],[23]. This enhanced potency, however, has to be used with care, as high densities of LNAs have been associated with hepatotoxicity.
A particularly elegant and widely used strategy is the gapmer architecture. Gapmers consist of a central stretch of DNA nucleotides flanked on both sides by chemically modified nucleotides such as MOE or LNA.
The modified flanks protect the ASO from nuclease degradation and stabilize the oligonucleotide, while the central “gap” of DNA is recognized by RNase H1, enabling catalytic cleavage of the target RNA. This dual functionality makes gapmers highly versatile: they combine the durability of chemical modifications with the enzymatic activity of the cell, making them one of the cornerstones of antisense therapeutics [24]. Other designs, such as mixmers or fully modified steric blockers, also rely on chemical modifications, but gapmers remain the most widely used architecture because they best illustrate how chemical design translates into biological activity [25].
Overall, chemical modifications define the therapeutic performance of ASOs. They extend stability, improve potency, and enhance biodistribution, but they can also introduce toxicities if used excessively, unevenly, or in repetitive motifs. Understanding and balancing these trade-offs is essential for rational ASO design [23-27]. In our project, these modification patterns are not just chemical details; they are integrated into our computational model as measurable features. A more detailed explanation of how modification-related parameters are quantified can be found in the model section of our project.
See Model - Modifications feature
Even with careful sequence design and chemical modifications, ASOs can sometimes cause problems by binding unintended RNAs or interacting with proteins in harmful ways. One of the most concerning side effects seen in clinical studies has been hepatotoxicity, where ASOs accumulate in the liver and trigger toxic responses [13]. Minimizing these risks is therefore a central challenge in bringing antisense therapies safely to patients.
To address this issue, researchers have recently proposed a novel safety strategy known as the BROTHERS (BRace On a THERapeutic aSo) system [16]. In this approach, the ASO is initially paired with a complementary peptide nucleic acid (PNA) strand that functions as a molecular brace. While bound to the Brother, the ASO remains inactive and cannot hybridize with RNA. Only in the presence of the correct RNA target does a toehold-mediated strand displacement reaction occur, releasing the ASO from its brace and allowing it to act. This selective activation occurs because the ASO has a higher binding affinity for its RNA target than for the PNA brace, ensuring that it is released only when the correct target is present.
This conditional design gives ASOs context-dependent activity: in cells that lack the target RNA, the ASO stays “off,” reducing the chance of off-target interactions and toxicity. In cells where the target RNA is abundant, the ASO is activated locally and carries out its intended function.
Even with advanced chemical modifications and safety systems like BROTHERS, ASOs may still affect healthy cells. To achieve true tumor selectivity, an additional layer of precision is needed. In our project, we address this challenge by applying the principle of synthetic lethality (SL) [26], which will be described in detail in the following section.
2. Synthetic Lethality
By targeting genes that become essential only when a cancer-specific mutation is present, we can selectively kill cancer cells while sparing healthy tissue. Unlike single-mutation drugs, this synthetic-lethality approach allows us to simultaneously cover multiple mutations within a heterogeneous tumor, reducing resistance and broadening patient eligibility.
In recent years, synthetic lethality has emerged as an important theme in the field of targeted cancer therapy, offering significant potential for the development of drugs that selectively eliminate cancer cells while preserving healthy tissue [27].
The original concept of Synthetic lethality (SL) is based on the simultaneous occurrence of abnormalities in the expression of two or more separate genes, including mutation, overexpression, or gene inhibition, which leads to cell death [28]. In the context of cancer, synthetic lethality utilizes the vulnerability of cancer cells caused by mutations in specific genes that make them dependent on alternative biological pathways. By targeting the gene essential to these pathways, synthetic lethality can induce cell death in cancer cells whereas, non-cancerous cells lacking the mutation retain functional primary pathways and are therefore preserved [29].
A key advantage of synthetic lethality is its ability to address mutations in genes that are otherwise considered "undruggable" or inaccessible by targeting a separate gene on which the cancer cell has become dependent [29].
Another major benefit is the potential to target a wide range of different mutations that converge on a common vulnerability, by inhibiting a single shared partner gene [30].
Although promising, synthetic lethality is still a relatively new and evolving field, and to date, only a few FDA-approved anti-cancer drugs have been developed based on this approach. All currently approved SL-based drugs—such as olaparib and some other PARP inhibitors target the same synthetic lethal interaction between BRCA1/2 mutations and PARP1 inhibition, used in the treatment of ovarian and breast cancers [31]. This highlights both the potential and the current limitations of the field, underscoring the need for further research to identify new synthetic lethal gene pairs for broader clinical applications.
3. Antibody-Oligonucleotide Conjugates (AOCs)
Our targeting vehicle. By conjugating the ASO and epitope to a monoclonal antibody, we ensure delivery directly to cancer cells, improving efficacy and reducing side effects. This precision targeting maximizes therapeutic impact while minimizing harm to healthy tissues.
Antibody–drug conjugates (ADCs) have transformed oncology by coupling the specificity of antibodies with the potency of cytotoxic small molecules. This strategy enabled the clinical use of payloads that were otherwise too toxic as standalone agents, highlighting antibodies as ideal delivery vehicles due to their specificity, long half-life, and favorable pharmacokinetics. Building on this principle, antibody-oligonucleotide conjugates (AOCs) represent the next step in targeted therapeutics. By replacing cytotoxic drugs with antisense oligonucleotides (ASOs), AOCs combine the precision of gene-silencing molecules with the targeted delivery capacity of antibodies, offering a safer and more selective therapeutic modality [32].
Even though oligonucleotides provide unmatched target selectivity at the genetic level, they face major delivery challenges, including short serum stability, poor membrane permeability, and lack of tissue specificity [32]. To overcome these challenges, we have selected Antibody-Oligonucleotide Conjugates (AOCs) as our preferred delivery method.
One of the most critical limitations in ASO therapy is off-target distribution, which reduces efficacy and increases toxicity. AOCs overcome this by utilizing monoclonal antibodies that recognize and bind to specific cell-surface receptors uniquely or abundantly expressed in disease-relevant tissues. Upon binding, the AOC undergoes receptor-mediated endocytosis, ensuring selective uptake into only the desired cells [32].
The mechanism of AOC uptake mimics a natural cellular process, receptor-mediated endocytosis, where binding of the antibody to its receptor triggers internalization of the entire AOC complex. Once inside the endosome, appropriate linker design enables release of the ASO into the cytoplasm where it can engage with its RNA target [32].
Another strength of the AOC platform is its modular design. The ASO remains constant, while only the antibody portion needs to be swapped to target a different tissue. This is advantageous not only for adapting to other diseases but also for customizing therapies based on patient-specific expression profiles (precision medicine).
AOC-based therapies are gaining rapid clinical traction. Companies such as Avidity Biosciences and Dyne Therapeutics have developed AOC platforms targeting muscular diseases (e.g.,DM1, FSHD, and DMD), with several programs already in Phase I/II clinical trials [33], [34]. This growing body of evidence supports AOCs as a translatable, clinically relevant, and innovative delivery method, especially for gene-targeting therapies where cell-type selectivity.
In our project, ONCOLIGO, we apply this approach by designing ASOs that silence cancer-driving genes in lung cancer and conjugating them to antibodies that selectively bind receptors overexpressed on lung tumor cells. This strategy maximizes therapeutic precision while minimizing systemic toxicity, paving the way for a novel and safer cancer treatment.
4. Epitope
A small, recognizable molecular tag that can be presented on the surface of cells - critical for guiding immune recognition or specific targeting.
The immune system works on identifying molecular signatures on pathogens or abnormal cells to trigger an immune response. Among its components, T cells are a type of white blood cells and can recognize tumor antigens that are generated by mutations in tumor cells. Some tumor antigens that are called tumor-associated antigens, are expressed in normal tissues as well but over-expressed in tumor cells. A different type of tumor antigens are tumor-specific neoantigens that occur due to somatic mutations that change the amino acid sequence, creating mutated peptides. The specific part in neoantigen that is recognized by the T cells are short peptides called neoepitope [35], [36].
Neoantigens are presented on the surface of tumor cells by major histocompatibility complex (MHC) molecules. The resulting peptide-MHC complex is specifically recognized by T cell receptors (TCRs) on T cells, which triggers activation of the cytotoxic immune response against the cancer cell (Figure 8). Those mutations that create the neoantigens do not appear in normal tissues, which make them ideal targets for T-cell immunotherapy [36], [37].
Cancer cells are known to evolve resistance to immunity against cancer, and those escape mechanisms can be fought using cancer immunotherapy. Increasing the amount of neoantigens can help enhance tumor-specific immune responses and give additional treatment to create an immunological memory in case of disease recurrence [37].
Epitopes are the specific regions of an antigen that interact with the immune system. B cells recognize epitopes through antibodies, while T cells recognize short peptide epitopes presented by MHC molecules. Because epitopes define the exact sites of immune recognition, they play a central role in the design of vaccines, antibody-based therapies, and T cell-based immunotherapies. In cancer, neoepitopes generated by tumor mutations represent ideal targets because they can be selectively recognized by T cells without affecting healthy tissues [35], [36], [37].
Cancer immunotherapy builds on the principle of using immunological methods to target an abnormal immune state and changing or adjusting the immune system's function. The goal in this kind of therapy is to enhance the ability of the body to recognize and target cancer cells and to eliminate those cells [38]. The main mechanism is recognition by cytotoxic T lymphocytes (CTLs) which are the responsible T cells for cancer elimination. In personalized therapeutic cancer vaccines that are designed to deliver neoantigens to trigger the immune response, this recognition by CTLs approach has been popular in recent years and may be a promising way in immunotherapy [38], [39], [40].
A major challenge in translating this knowledge into therapies is identifying which mutations will generate immunogenic epitopes capable of binding to MHC molecules and activating T cells. Tumors often contain thousands of mutations, making experimental testing of all candidates impractical. To address this, computational algorithms have been developed to predict which mutated peptides are most likely to function as neoepitopes. These tools combine data on peptide-MHC binding affinity, processing, presentation, and predicted immunogenicity, helping to prioritize the most promising targets for therapeutic development [35], [36], [38].
A key principle in cancer immunotherapy is not only to generate new immune responses but also to amplify the activity of existing CTLs. These T cells are already capable of recognizing abnormal or foreign epitopes presented on MHC class I molecules. By delivering tumor-specific epitopes through therapeutic cancer vaccines, it is possible to enhance the presentation of these antigens and boost CTL activation. In this way, vaccines help direct the immune system toward tumor cells with greater strength and specificity, increasing the likelihood of effective tumor clearance [36], [37], [38].
Building on this concept, epitopes can be delivered using antibody-based systems. Antibodies provide highly selective recognition of tumor-associated surface proteins and can be engineered to carry epitope peptides as a functional payload. Once internalized into tumor cells, these peptides can be processed and presented on MHC molecules, thereby converting the cancer cell into a direct target for epitope-specific CTLs. Antibody-mediated epitope delivery offers an advantage over vaccines alone by coupling tumor targeting with immediate T cell recognition, creating a more focused and potent immune attack [39].
5. Computational Model
Our model aims to design and prioritize highly effective ASO sequences by integrating novel predictive features and synthetic lethality principles, ensuring precise cancer cell targeting. In parallel, it guides antibody–epitope optimization to enhance delivery, immune engagement, and therapeutic impact.
See our Model and Software pages for details.
Our platform overcomes the limitations of existing ASO design tools through a comprehensive, data-driven approach. Learn more in our Comparison To Competitors.
For more details, see our Engineering page.
Alongside our cancer-focused experiments, we are also exploring the use of ASOs in Saccharomyces cerevisiae. To date, ASO-based gene knockdown has been rarely tested in yeast. By combining our computational design pipeline with this well-established model organism, we aim to create a simple, scalable platform for targeted gene knockdown in yeast. This will not only validate the versatility of our ASO design strategy but also provide a valuable tool for the broader research community, as yeast remains one of the most extensively studied eukaryotic models.
Our short-term objective within the iGEM competition is to generate clear in vitro results through iterative cycles of model prediction and experimental validation. By doing so, we aim to demonstrate the feasibility of our modular strategy and establish a reliable computational–experimental pipeline. In parallel, we are developing a robust and scalable ASO design model that can accurately predict effective candidates and minimize off-target risks.
Beyond this immediate focus, we also aim to highlight the versatility of our platform by extending ASO-based knockdown into Saccharomyces cerevisiae. Success in this model organism would showcase the adaptability of our design engine beyond cancer, opening opportunities in functional genomics and systems biology.
Looking further ahead, our long-term vision is to advance this platform into a clinically viable therapy for non-small cell lung cancer (NSCLC), with the ambition of reaching a first-in-human treatment within the next decade. Alongside this therapeutic development, we seek to establish a robust ASO design model that can be licensed and adapted beyond lung cancer, ultimately enabling applications across other tumor types and even genetic diseases more broadly.
Our inspiration for this project comes directly from the urgent unmet need in lung cancer treatment. During our meeting with Prof. Amir Onn, Chair of the Institute of Pulmonary Oncology at Sheba Medical Center, he shared the story of a young mother he has been treating since 2016. After nearly a decade of battling lung cancer, all available therapies had failed. “I have nothing left to give her,” he told us. “No drug works anymore.”
We also drew inspiration from conversations with patients themselves. As part of our Human Practices work, we interviewed Ellen Nemetz, a 64-year-old artist from Arizona diagnosed with stage IV lung cancer carrying the rare mutation. Ellen described the overwhelming physical burden of the disease, from painful paraneoplastic symptoms to the harsh side effects of chemotherapy and immunotherapy. Despite multiple treatment lines, the cancer continued to progress, leaving her with limited options.
These stories highlight the human urgency behind our project and remind us that beyond statistics, each patient’s journey is unique and deeply personal.
To read more about these stories, visit our Human Practices page.
