DeciBio’s Post

Struggling to push the boundaries of personalized medicine and immunotherapy? Neoantigens might be the breakthrough you've been searching for. Julie Gil, PhD, Principal Solution Scientist at Velsera, discusses the cutting-edge world of neoantigen discovery—an innovation transforming cancer treatment and personalized medicine. In this presentation, Julie reveals how advances in bioinformatics, machine learning, and genomics are unlocking new possibilities for patient care. Explore the future of immunotherapy! Watch Julie’s insightful presentation on-demand now and discover how neoantigens could hold the key to more effective, personalized treatments. https://lnkd.in/grW258ZA   #PersonalizedMedicine #Neoantigens #EraOfVelsera 

Terrific panel discussion at #WorldCDx on Digital Biomarkers & Spatial Biology to Revolutionize Precision Medicine. It was great to share more about Friends of Cancer Research Digital and Computational Pathology Harmonization Project evaluating variability in biomarker assessments across platforms to help support alignment in the field. A few takeaways from our discussion:   - Appropriate validation and transparency of tool performance to determine if the tools are fit-for-purpose will be critical. Defining “ground truth” continues to be a challenge with novel biomarkers and the variability observed with pathologists. - Robust reference sets will be key- detailing input parameters (staining, scanning, methodologies), clinical/specimen characteristics, and outcomes. This will require alignment on data dictionaries and methods. - This field will only get more complex- there is a lot of promise but still a lot of questions! So great to see so many of our project partners at the conference! Learn more about our work here: https://lnkd.in/eqx5iYyr

The ELLKAY User Group meeting is right around the corner. Hear from our clients first hand the impact that ELLKAY solutions have had on their bottom line. #teamellkay #pathology

⭐ Join us on October 10th at 12 pm ET for our AI4C Lecture Series, featuring William Hsu, Ph.D., and hosted by Katherine Andriole, Ph.D., FACR, FSIIM, FSPIE. Our growing capacity to collect vast amounts of longitudinal, multimodal data must be matched with a comparable ability to extract biological insights and actionable information. This talk describes Dr. Hsu's lab efforts to develop and validate computational tools that combine clinical, diagnostic imaging, and molecular data to help physicians formulate timely, accurate, and personalized management strategies for individual patients, particularly in the context of cancer screening. Register in advance for this webinar: https://lnkd.in/eKxQ_Uu8 #ArtificialIntelligence #CancerScreening

Learn about how AI is being developed to help create personalized cancer screening strategies with the AI4C Lecture Series!

The described cBioportal features for joint longitudinal clinicogenomic data analysis are also available to explore the recently published MSK-CHORD data (see https://lnkd.in/dX8KxV9R). MSK-CHORD was made possible by leveraging the manually curated AACR Project GENIE BPC data to train NLP models and expand it to 24K MSK-IMPACT cases

How can you use rocket science in the field of healthcare? Probably a question no one has ever asked, right? Here at Calliope Biophysics, we excel in asking questions no one in their right mind would think of, and then use our novel approaches to make a difference in people's lives. The question above isn't rhetorical. Our first product in development does use rocket science in healthcare! It gives real time data to a surgeon so they can know where cancer is as before they remove it, making it so healthy tissue isn't removed needlessly. We try to ask the unique questions here, and hope to deliver on some extraordinary products. Follow us for updates as we continue development and revolutionize cancer detection for the better.

Feature selection plays a crucial role in the analysis and Machine Learning applications using high-dimensional biomedical data, enhancing model interpretability, reducing computational complexity, and improving predictive performance. In domains like breast cancer diagnosis and diabetes progression prediction, identifying key features is essential for creating accurate and interpretable models. Traditional methods like Lasso, mutual information-based selection, and ANOVA F-value selection have shown success in various applications. However, they often struggle with stability and performance across diverse biomedical datasets. Introducing Regularized Sparse Optimization (RSO), a custom adaptive feature selection method that combines the strengths of L1 and L2 regularization in a flexible, data-driven objective function. RSO aims to address the challenges faced by traditional methods, offering a more robust approach to feature selection in biomedical data analysis. #FeatureSelection #BiomedicalData #DataAnalysis #RSO #Innovation

🌟 Exciting Lecture at the 22nd Israeli Bioinformatics Symposium! 🌟 I had the honor of presenting at IBS_2024, where I discussed our groundbreaking research on identifying the tumor microbiome and its utilization as a diagnostic tool. Here’s a quick overview of what we explored: 🔍 Key Points from the Lecture: 1. Tumor Microbiome Profiling: We analyzed microbiome composition within cancer tumors using whole genome and transcriptome sequencing. 2. Machine Learning for Tissue Discrimination: Developed a supervised ML model to distinguish between "Normal" and "Tumor" tissue based on microbial profiles. 3. Molecular Pathway Analysis: Investigated the interactions between hazardous microbes and tumor tissue characteristics to understand their role in cancer progression. Our study underscores the power of bioinformatics and machine learning in advancing cancer diagnostics through microbial analysis. #Bioinformatics #CancerResearch #MachineLearning #IBS2024

We are thrilled to share the presentation of Stephanie Chrysanthou, PhD, our R&D Manager at Integrated Genomics Operation Memorial Sloan Kettering Cancer Center, on detection of germline alterations in solid #cancer patient samples by adaptive sampling 🙌 presented at the Nanopore Community Meeting conference in Boston! Our core has closely partnered with the Molecular Pathology & Diagnostics lab to benchmark Adaptive Sampling long read #sequencing for detecting germline #mutations against prior MSK-IMPACT profiles generated by hybridization capture short-read sequencing. Adaptive Sampling is a promising approach due to reduced technical complexity and TAT, there is no need for sample batching and it facilitates the resolution of large complex structural #variants. Due to the live #basecalling capacity of the #nanopore technology, the sequencing can be stopped if a fragment does not match the uploaded target .bed file, thus enriching on-target fragment #sequencing. Briefly, our on-instrument data workflow includes: 1) filtering of the raw .pod5 files based on the on-target decision, 2) SUP basecalling, 3) GRCh38 alignment, 4) SNV/SV calling using #EPI2ME and #SEVERUS pipelines.  On a routine basis, this workflow, from #hmwDNA extraction to data analysis, takes 3.5 days. Feel free to watch the full presentation here 👇 (it's free to register and you'll get access to all ONT conference presentations!) Oxford Nanopore Technologies https://lnkd.in/ezE_vsYT