Deepcell’s Post

🔍 How can simply looking at cells through the Deepcell lens guide cancer therapy decisions? In collaboration with AbbVie (data presented at #AACR2024), we demonstrated direct linking between cell morphology to gene expression and function in a lung cancer patient tissue biopsy.   🔬 We identified a rare malignant subpopulation (Cluster 0) with a drastically different gene expression profile compared to the majority of the patient's cancer cells (Cluster 2). 💡 Through sorting, we uncovered key differences in gene expression and pathways, including upregulation of STAT3 (which promotes tumor growth, survival, angiogenesis, immune evasion, and inflammation) and NF-kB signaling (fueling inflammation, tumor survival, resistance to apoptosis, and metastasis) in this rare subpopulation. 🎯 These insights suggest therapeutically targeting STAT3 and NF-kB signaling to effectively combat Cluster 0. The results underscore #REM-I’s potential to unveil hidden cellular subtypes, drive groundbreaking cancer therapies, and transform patient outcomes. Start your journey to harness the power of AI for your application by requesting a demo here: https://lnkd.in/g_nPdRR4 Link to AACR poster here: https://lnkd.in/gF9zA-T4 #PrecisionOncology #CancerResearch #Deepcell #AACR2024 #CancerTherapies #Morpholomics

Congratulations to Deepcell’s scientific cofounder Prof. Euan Ashley for being appointed as the chair of Stanford Department of Medicine.

It is a pleasure to see groundbreaking applications of Deepcell's AI technology highlighted in the discussion organized by Rotterdam Square led by Prof. Peter J. van der Spek. From detecting small numbers of affected cells in blood and urine to enhancing diagnostics across fields like cardiology, oncology, and immunology, Deepcell's AI-enabled solutions continue to push the boundaries of personalized healthcare. "In the time it takes a pathologist to view ten cells and drink a cup of coffee, her AI-enabled colleague can scan up to 100,000 cells and compare them with an extensive library. " Interested to understand how our groundbreaking products can transform your clinical workflows? Reach out to us and request a demo: https://lnkd.in/g_nPdRR4

Truly, the natural world hath never been more within our grasp!

🔬 Unlocking the Complexity of Cell Death with High-Dimensional Morphology Data Using the REM-I platform, we've captured the intricate and dynamic process of cell death through high-dimensional morphology data. REM-I’s discriminative resolution allows us to identify subtle morphological shifts in the very early and mid-phases of apoptosis. By leveraging deep learning and computer vision, the platform provides orthogonal biological insights that not only complement but often surpass what traditional cytotechnology and protein-based biomarkers can reveal. Check out the visualizations below: (A) Morphology UMAP: Generated by REM-I and colored by cluster using the Leiden algorithm, with representative images included. These morphotype clusters align with morphological changes as cells undergo apoptosis and necrosis. (B) UMAP Visualization: Colored by population density, this visualization demonstrates how different treatments result in varying morphological expressions. Representative images from each treatment condition are displayed below. #Morpholomics #DeepLearning #ComputerVision #CellBiology #Apoptosis #Morphology #Deepcell

Happy olympics season! #Olympics2024

🧑🔬 In a new preprint led by investigators at the University of Zurich, melanoma researchers distinguished between melanocytic and mesenchymal cell phenotypes. 🔬 Using label-free high-dimensional morphology profiling, melanoma cell lines were imaged on the REM-I and resulting 115 deep learning and computer vision features were visualized as a reduced dimensionality projection (Figure A). 🔍 Distinct morphotype groups emerged, with melanocytic and mesenchymal cells localizing with respect to a combination of deep learning and morphometric features related to cell texture and size (Figure B-C). 💡 Notably, the REM-I system detected and discerned cells with high pigmentation levels from other melanocytic cells. 🌟 These findings are promising for future applications such as stratifying patient samples based on cell type (Figure B). Check out the preprint available at bioRxiv, titled ”Label-Free Melanoma Phenotype Classification Using Artificial Intelligence-Based Morphological Profiling”. 📑