URI Pharmaceutical Development Institute’s Post

The synergy of human health and space technology is as old as space exploration itself, and there is huge potential for applying the lessons learned in space for commercial success on Earth. It's great to see more support for Scottish start-ups in this important and growing sector:

#DAY13 🚀 Excited to share insights on Day 13 of my 100 Days Technical Transformation Challenge! 🚀 Join me on a journey through the complexities of linear regression, where we explore dimensionality reduction and feature selection. 📈 In this concise guide, we unravel the advantages, disadvantages, and real-world applications of each technique. 1️⃣ Dimensionality Reduction: 🌐 Advantages: Enhance computational efficiency and mitigate overfitting. Disadvantages: Navigate the trade-off between interpretability and information loss. Use Cases: From image processing to genomics, discover the transformative power of reducing dataset dimensions. 2️⃣ Feature Selection: 🎯 Advantages: Improve model performance and boost interpretability. Disadvantages: Mindful of potential overlooking of interactions and method sensitivity. Use Cases: Dive into biomedical research and financial modeling, where selecting impactful features is key. 3️⃣ Backward Elimination: ⏪ Advantages: Systematically refine models and uphold statistical rigor. Disadvantages: Address assumptions of linearity and potential omission of interactions. Use Cases: Explore economic modeling, identifying influential factors shaping economic variables. 4️⃣ Wrapper Methods: 🔄 Advantages: Tailor feature selection to specific models and consider feature interactions. Disadvantages: Navigate computational intensity and model dependency. Use Cases: Uncover the power of wrapper methods in medical diagnosis for maximizing predictive performance. 🔗 Ready to explore these techniques and transform your technical prowess? Dive into the full guide now! 📚 #TechnicalTransformation #DataScience #LinearRegression #FeatureSelection #ChallengeDay13 🔗 What's your experience with dimensionality reduction and feature selection? Share your insights or join the conversation! 💬 Let's learn and grow together in this 100 Days Technical Transformation Challenge! 🚀 #TechCommunity #LearningJourney Check_Out_Detailed_blog:-https://lnkd.in/gAWwxtt8

Did you know the Society for Imaging Informatics in Medicine (SIIM) Hackathon isn't just for coders? Join me next week in an informative webinar to talk about the past, present and future of SIIM Hackathons. As part of this webinar, we'll also explore the impact this hackathon has had on the industry over the past 10 years (yes, hard to believe it has been 10 years). https://lnkd.in/g8pCcVCj

How are #ADRD researchers leveraging shared data and resources to advance discoveries in #Alzheimer’s research? Arthur Toga and Niranjan Bose led a great scientific session yesterday at #AAIC24 exploring this question – here are some highlights ⬇️   Dr. Nicolai Franzmeier demonstrated the benefits of cloud-based platforms like AD Workbench, which can help researchers include more contributors across different environments, create a transparent and reproducible environment, and access multiple data sets directly.    Ganna Blazhenets discussed how her team used the Global Alzheimer's Association Interactive Network platform to do comparisons of the real-world centiloid values that they generated as part of the IDEAS study across five large research datasets that can be accessed through GAAIN and Laboratory of Neuro Imaging (LONI).   Tessa Harrison discussed the value of being able to harmonize PET data across sites and studies. The Berkeley PET Imaging Pipeline has been used to process over 15,000 PET scans, allowing researchers to combine cohorts and create more representative samples.    Ioannis Pappas presented a case study on using the GAAIN and AD Workbench platforms to analyze the Baltimore Longitudinal Study of Aging, creating a unified framework for federated discovery, interrogation, and analysis.

“Apps to close gaps” Earlier this month, we wrapped Apollo cohort 1 at the SDA TAP Lab. Under Sean Allen’s leadership, the team of multiple companies and universities set out to help Space Force operators make decisions within increasingly complex space terrain, replace their sticky notes with autonomous applications and ultimately, avoid operational surprise… And we had to go fast. Within 49 days, our team integrated applications alongside several companies to quickly detect and characterize the start of kill chains using machine learning and modern visualization tools, reducing processes from hours to seconds. The interoperable app model reduces cognitive load on operators from managing multiple systems just as much as the underlying tech. And these aren’t just talking points…we built and demonstrated that it can be done! Today's product-market-fit is both about the product being delivered AND the company delivering them. The government wants Silicon Valley/SaaS-style products. The question is, can they attract and sustain the types of companies capable of producing them? Katalyst participated in the TAP Lab for two reasons. First, to work a mission that matters: supporting US space superiority. Second, we wanted to answer three questions: 1. Is our product nice-to-have or must-have? 2. How much is that worth? 3. Can the government acquire software applications via license and integrate them into operations on relevant timelines? The SDA TAP Lab has brought us closer to answering these questions than ever before. I see the role of the SDA TAP Lab prove there is a real model, with a real acquisition pathway, and with real money behind it. The first success will pave the way for more, encouraging a wave of companies to develop even better tools for the future. Looking forward to Apollo Cohort 2! #sda #space #ussf

Interesting announcement by Singular Genomics. Adding spatial biology capabilities to their G4 sequencer with an upgrade, and a new instrument called the G4x that does NGS and spatial in situ sequencing for RNA and potentially for protein barcodes plus H&E equivalent staining. The total area that can be imaged is an impressive 40cm2 across all the flow cells (I think). Available by EOY. The gene multiplexy seems low but the number of samples that can be analyzed in parallel is phenomenal. If this works as Singular claims it could have a negative impact on Nanostring (filed for Chapter 11 protection) and 10x Genomics spatial systems. Recollect in situ sequencing being done on the Life Tech (now Thermo) SOLiD Seq by Hyb NGS system and on the Illumina Seq by Synth HiSeq, in collaboration with George Church's lab about 10+ years ago.

🔍 Explore the future of medical predictions with our Prognostic Meta-Analysis Series! 🌐📊 Delve into the art of synthesizing prognostic studies, unraveling insights into predictive modeling, and understanding the dynamics of meta-analytic approaches. 🧬 Join leading experts as they navigate the realm of prognostic research, sharing valuable knowledge and unveiling the latest advancements. Elevate your understanding of predictive analytics and be at the forefront of groundbreaking discoveries! 🚀💡 Join our Grade 1 series. The last date of registration is 15th February 2024. Grade 1 is *compulsory* for attending advanced grades. https://lnkd.in/d5835VN6 #PrognosticAnalysis #ResearchInnovation #FuturePredictions

Minnesota Emerging Software Advisory (MESA) and Jim Moar have been excellent guides to our leadership throughout the growth process. Here is a great post on how they helped us on our path to success! #entrepreneurship #tech #ai #startup

Reproducibility is an essential principle underpinning all scientific research, but it is challenging to achieve in a host of disciplines, including neuroimaging. A new paper published in Nature Methods (add link), led by #UQEECS researchers, explores the exciting possibilities behind Neurodesk which offers a suite of neuroimaging software, accessible through a browser-based interface. Neurodesk marks a significant step towards revolutionising neuroimaging data analysis. By overcoming technical challenges and offering a user-friendly interface, Neurodesk sets a precedent for accessible research environments across scientific disciplines, promising a more reproducible and collaborative future in data analysis. The Neurodesk platform is built by and for the scientific community to enable anyone to perform reproducible analyses with ease. Learn more: https://lnkd.in/gGEFRD7m