To predict and effectively treat patients most likely to benefit from specific therapies, our precision medicine platform integrates advanced machine learning with comprehensive patient data analysis. We start by curating and normalizing a large database that includes detailed disease status and severity information, alongside normal controls. This data is used to identify key biomarkers through the use of RNA sequencing and meta-analysis, focusing on PDE4-effector genes. By leveraging systems biology, we confirm these biomarkers as critical drivers of disease pathology. Machine learning then refines the selection of biomarkers, assessing their impact on disease activity and severity, and optimizes the number of key biomarkers needed to accurately predict patient responses to PDE4 inhibitors. This approach ensures targeted and effective treatments, enhancing the likelihood of positive patient outcomes and remission. As part of our ongoing work towards a precision medicine test to identify patients likely to respond to PALI-2108, RNAseq datasets were analyzed using a standardized bioinformatics pipeline, which generates normalized TPM counts for each gene. The analysis found that PDE4B expression is significantly higher in most colitis patients, encompassing both adult and pediatric cohorts. This suggests that PDE4B could be a crucial factor in patient stratification. In adults, a PDE4B over-expression level was determined that the Company believes accurately identifies 70% of moderate to severe cases. In pediatric patients, a PDE4B over-expression level was determined that the Company believes accurately identifies 90% of those with moderate to severe colitis.

We have used well accepted preclinical models to confirm this approach. We have demonstrated increased colon tissue PDE4B levels with induction of colitis and significant reductions in colon tissue PDE4B expression in response to increasing doses of PALI-2108 in an accepted mouse model of colitis. We also demonstrated, closely correlated and increasing levels of colon tissue cAMP, and closely correlated to colon tissue inflammatory cytokines such as TNF-alpha. We additionally observed a significant reduction in disease activity index score over time compared to the control treated group. Further, body weight loss and changes in colon length were attenuated in increasing dosage groups, showcasing the potential for targeted efficacy.

To predict and effectively treat patients most likely to benefit from specific therapies, our precision medicine platform integrates advanced machine learning with comprehensive patient data analysis. We start by curating and normalizing a large database that includes detailed disease status and severity information, alongside normal controls. This data is used to identify key biomarkers through the use of RNA sequencing and meta-analysis, focusing on PDE4-effector genes. By leveraging systems biology, we confirm these biomarkers as critical drivers of disease pathology. Machine learning then refines the selection of biomarkers, assessing their impact on disease activity and severity, and optimizes the number of key biomarkers needed to accurately predict patient responses to PDE4 inhibitors. This approach ensures targeted and effective treatments, enhancing the likelihood of positive patient outcomes and remission. As part of our ongoing work towards a precision medicine test to identify patients likely to respond to PALI-2108, RNAseq datasets were analyzed using a standardized bioinformatics pipeline, which generates normalized TPM counts for each gene. The analysis found that PDE4B expression is significantly higher in most colitis patients, encompassing both adult and pediatric cohorts. This suggests that PDE4B could be a crucial factor in patient stratification. In adults, a PDE4B over-expression level was determined that the Company believes accurately identifies 70% of moderate to severe cases. In pediatric patients, a PDE4B over-expression level was determined that the Company believes accurately identifies 90% of those with moderate to severe colitis.

We have used well accepted preclinical models to confirm this approach. We have demonstrated increased colon tissue PDE4B levels with induction of colitis and significant reductions in colon tissue PDE4B expression in response to increasing doses of PALI-2108 in an accepted mouse model of colitis. We also demonstrated, closely correlated and increasing levels of colon tissue cAMP, and closely correlated to colon tissue inflammatory cytokines such as TNF-alpha. We additionally observed a significant reduction in disease activity index score over time compared to the control treated group. Further, body weight loss and changes in colon length were attenuated in increasing dosage groups, showcasing the potential for targeted efficacy.

• Identified PDE4-effector genes predictive of patient response to PDE4 inhibitors using advanced bioinformatics tools.
• Selected key drivers of disease pathology for moderately to severely active UC patients through systems biology and machine learning.
• Ranked genes by their importance using meta-analysis, focusing on both pro-inflammatory and anti-inflammatory markers.
• Generated a weighted Disease Activity Score (DAS) for precise patient stratification and treatment targeting.