Christian A. Lang

Led by Mihai Pop, "Reference-guided assembly of metagenomes with MetaCompass" is out in Cell Reports Methods. The number of complete bacterial genome sequences available in public databases is rapidly increasing, but this resource has yet to be effectively used in guiding the analysis of microbiome samples. We aimed to develop an effective technique for using vast genome sequence collections to serve as a reference in the analysis of microbiome data, addressing the challenge of determining which, among the hundreds of thousands of genome sequence that are publicly available, can be used as references for guiding the analysis of a given sample. The software is released at https://lnkd.in/ejv5JGPK under the Artistic License 2.0.

9

Tamarind Bio Secures $13.6M Series A to Make AI More Accessible for Biology Tamarind’s platform for model coordination and inference enables user-friendly AI tools for life science researchers. Sherry Liu, Deniz Kavi #ArtificialIntelligence #biology #DrugDiscovery https://hubs.li/Q044vB360

18

1 Comment

Outliers and missing data in #omics datasets can complicate #biocomputational analysis – but they can also reveal unexpected insights. In this article, Sapient data scientists Shreya Gaddam and Cameron Lamoureux, PhD explore practical approaches to identify and address missingness and outliers in large-scale biological data, and how thoughtful analytical strategies can improve the reliability of downstream results. If you're working with high-dimensional datasets like those generated by discovery #proteomics and #metabolomics studies, this is surely worth a read. 🔗 https://lnkd.in/g6tSCVZ4 #datascience #mindthegaps #biologicaldata

12

Thrilled to announce the official release of SDialog ! SDialog is an open-source Python toolkit for synthetic dialogue generation, evaluation, and analysis, designed to make creating high-quality, controllable dialogue data easier and more scalable. This toolkit has been a big part of our teamwork during JSALT workshop, and I’m really excited to see it go public! I actively participated in the development of the clinical reason for visits to inform the dialogue, the clinical dialog physical context, the clinical evaluation and the LLM as a judge which measure the clinical integrity of the generate dialogue, analyze factuality, omission, safety, communication, and cultural adaptation, and score the conversational AI from clinical depth standpoint. 👀 🔍 Key Features Key features of SDialog: - Persona-Driven Simulation: Easily create multi-agent simulations with rich personas and contexts. - Composable Orchestration: Precisely control dialogue flow and agent behavior with simple, pipe-like composition. - Built-in Evaluation: Use a suite of metrics and LLM-as-judge capabilities to evaluate and compare dialogue quality. - Mechanistic Interpretability: Go beyond black-box generation. Inspect and even steer model activations to understand and influence behavior. - LLM Interoperability: Works out-of-the-box with backends like OpenAI, HuggingFace, Ollama, AWS Bedrock, and more. I’m very proud of this release, and especially of the clinical evaluation tools. 🔗 Check it out on GitHub (Idiap Research Institute): https://lnkd.in/eJwiXNZb Give it a ⭐️ if you find it interesting! A big thanks to Sergio Burdisso, Yanis Labrak, Séverin BAROUDI, Thomas Schaaf, Andrew Perrault, Amy Chun, and all the others. Also, special thanks to Ricard Marxer & Sanjeev Khudanpur for giving me the opportunity to participate to JSALT25 !

28

𝗧𝗵𝗲 𝗲𝗿𝗮 𝗼𝗳 𝘀𝗶𝗹𝗼𝗲𝗱 𝗔𝗜 𝗺𝗼𝗱𝗲𝗹𝘀 𝗶𝗻 𝗰𝗮𝗻𝗰𝗲𝗿 𝗿𝗲𝘀𝗲𝗮𝗿𝗰𝗵 𝗶𝘀 𝗼𝘃𝗲𝗿. The transition from classical machine learning to 𝗙𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻 𝗠𝗼𝗱𝗲𝗹𝘀 (𝗙𝗠𝘀) marks a fundamental shift in how we approach precision medicine. Integrating genomics, proteomics, imaging, and clinical data is no longer optional. It is the only way to unravel the complexity of cancer. I am thrilled to share that our latest research, "𝗙𝗿𝗼𝗺 𝗰𝗹𝗮𝘀𝘀𝗶𝗰𝗮𝗹 𝗺𝗮𝗰𝗵𝗶𝗻𝗲 𝗹𝗲𝗮𝗿𝗻𝗶𝗻𝗴 𝘁𝗼 𝗲𝗺𝗲𝗿𝗴𝗶𝗻𝗴 𝗳𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻 𝗺𝗼𝗱𝗲𝗹𝘀: 𝗿𝗲𝘃𝗶𝗲𝘄 𝗼𝗻 𝗺𝘂𝗹𝘁𝗶𝗺𝗼𝗱𝗮𝗹 𝗱𝗮𝘁𝗮 𝗶𝗻𝘁𝗲𝗴𝗿𝗮𝘁𝗶𝗼𝗻 𝗳𝗼𝗿 𝗰𝗮𝗻𝗰𝗲𝗿 𝗿𝗲𝘀𝗲𝗮𝗿𝗰𝗵", has officially been published in 𝗔𝗿𝘁𝗶𝗳𝗶𝗰𝗶𝗮𝗹 𝗜𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 𝗥𝗲𝘃𝗶𝗲𝘄. In this paper, we provide the first systematic map of the transition from conventional machine learning to advanced 𝗙𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻 𝗠𝗼𝗱𝗲𝗹𝘀 (𝗙𝗠𝘀). We explore: 🚀 The shift from task-specific models to large-scale pre-trained backbones. 🧬 Holistic strategies for multi-omics and advanced imaging integration. 🗺️ A roadmap for the forthcoming era of large-scale AI models in oncology. 🎯 Our goal is to provide researchers and clinicians with the essential groundwork to revolutionize cancer diagnosis, prognosis, and biomarker discovery. 📖 𝗙𝘂𝗹𝗹-𝘁𝗲𝘅𝘁 𝗔𝗰𝗰𝗲𝘀𝘀: 🔗 https://rdcu.be/e73oj 💻 𝗘𝘅𝗽𝗹𝗼𝗿𝗲 𝗼𝘂𝗿 𝗚𝗶𝘁𝗛𝘂𝗯 𝗿𝗲𝗽𝗼𝘀𝗶𝘁𝗼𝗿𝘆: https://lnkd.in/eGNbBBN9 A huge thank you to my incredible co-authors: Muhammad Waqas Sheikh, PhD, Maliazurina B. Saad, Eman Showkatian, Rukhmini Roy, Hui Xu, Wentao Li, Joe Chang, Zhongxing Liao, Cara Haymaker, Luisa M. Solis Soto, @Carol C. Wu @Natalie I. Vokes, Xiuning Le, @Lauren A. Byers @Don Gibbons @John V. Heymach, Jianjun Zhang, Jia Wu #AI #MachineLearning #CancerResearch #FoundationModels #DeepLearning #PrecisionMedicine #MultimodalData #AIinMedicine #SpringerNature #Oncology

47

4 Comments