JOIN US IN SHAPING THE CONNECTIONS BETWEEN CANCER RESEARCH AND AI.
Our mission is to enhance cancer treatment and knowledge by harnessing the power of computer intelligence. Guided by our vision that computers will eventually be able to think and perform similar tasks as humans, we strive to develop cutting-edge artificial intelligence and statistical approaches for building system models of tumors. This way opening the doors to collaboration with experimental labs and pharmaceutical companies . Our lab fosters a diverse, collaborative environment for advancing cancer research.
At Tumor-AI Lab, we develop artificial intelligence algorithms to advance basic science and cancer treatment. Our lab focuses on three areas:
Broadening the coverage of precision treatment through novel AI tools that transform insights from the transcriptomes of cancer cells into potential drugs and specific gene alterations targeting particular cancers.
Predicting emergent gene and protein traits by extracting multi-modal gene data through a large language model and knowledge injection techniques that combine structured and unstructured databases.
Deconstructing immune responses through imaging and analysis of single-cell and spatial biology.
Sept 2024: Access to world's most powerful supercomputer. With over 37,000 GPU, we are ready to tackle computational challenges that typical on-premise clusters or cloud cannot handle.
Sept 2024: Submitted two Manuscripts on Graphs. Congrats Mikaela, Luis, and Yue, and co-authors! Oct is coming!
Sept 2024: Our AI research at thewas recently featured on KOAT News! https://www.koat.com/article/new-mexico-unm-cancer-center-ai-research/62073232
Sept 2024: ScDist accepted to Nature Comm! Statistically principled approach for identifying pertubation in single cell data. https://www.nature.com/articles/s41467-024-51649-3
Aug 2024: LitGene, a transformer leveraging text knowledge for gene representations, predicting disease risk genes, grounding predictions to published literature. In bioArxiv now https://doi.org/10.1101/2024.08.07.606674 . Also the website https://litgene.avisahuai.com/
June 2024: Submitted Four Manuscripts for New Multimodal Approaches from TumorAI. Submitted four new manuscripts featuring innovative multimodal approaches DeepVul, MedGraphNet, Histopath-fusion, and Genen. Congratulations to Kushal, Ala, Macaulay, Yue, Kaibin, David, Luis, Mikaela, and Michael. Your hard work has truly paid off. Fingers crossed!
March 2024: Luis' Expert Guidance Propels Our High School Trainee Ganesh to Prestigious Awards! With Luis' mentorship, Ganesh topped in the Senior Division - Medicine & Health Sciences, also getting Donald Partridge Memorial Neuroscience and Richard Bild Memorial Research Challenge Awards.
March 2024: GeneLLM accepted at the ICLR 2024. A hearty congratulations to Ala, Macaulay, David, Kushal, and the rest!
February 2024: Dr Kushal Virupakshappa gets awarded the 2024 AAI Intersect Fellowship for Computational Scientists and Immunologists: Congratulations!
Jan 2024: Dr. Sarah Adams' project on DOD Ovarian Cancer Clinical Trials Academy has been funded! Our lab is thrilled to be part of the ACCELERATOR, an AI tool for ovarian cancer trials.
December 2023: Our New H100 Machine Running: Happy holidays!
November 2023: David Arredondo Joins the Lab as Postdoctoral Trainee: David earned his degree from the UNM School of Engineering, specializing in DNA Nanotechnology and Molecular computing
November 2023: Explore TumorAI Lab's Research in Avi's Presentation at UNM Grand Rounds.
See the presentation here
October 2023: BSGP Student Clara Bertoni Joins TumorAI Lab for Rotation in Single Cell Analysis.
New Pre-Print
We’re thrilled to announce the release of our latest preprint in collaboration with Rafa Irizarry's Lab! We’ve developed a principled approach, scDist, to tackle the challenge of individual and cohort variability in single-cell data. As the adoption of single-cell technologies grows, the importance of rigorous and principled approaches to handle the resulting data variation cannot be overstated. It’s crucial for improving the accuracy of our findings and reducing false positives. Adoption of this rigrous tool allowed us to detect robust perturbations in DC, pDC, and NK cells in COVID-19 and immunotherapy treatment responses.As single-cell datasets continue to expand, we believe that there is even more work to be done by those are passionate about the intersection of AI, statistics, and biology. With collective effort and open discussion, we can drive this field forward. Let’s make a difference together!
Take a look at our pre-print here.
Photo by National Cancer Institute on Unsplash.
Our Latest Publication in Cancer Discovery
This research publication in Cancer Discovery explores the discovery of bipotent drugs as a new solution for traditional combination therapies through BipotentR to find cancer cell-specific regulators that help modulate tumor immunity and a number of other oncogenic pathways.
This tool presents relative importance to evaluating patient response to treatment and the identification of new targets for drugs that allows for efficient elimination of tumors and cancerous cells. When this approach is taken and the topmost candidate tumors are inhibited through the suppression of their metabolism, the cancer cells are more likely to be killed/destroyed.
To learn more about this publication, click here.
If you are interested in some other of our publications, please refer to our page "Publications".
We believe in diversity and inclusivity for everyone, and that are dedicated to fostering these values inside our many cutting edge projects. A concept that translated directly to all that we do.
This is one of the reasons why we want to keep our tools and projects available to the public. All of our Artificial Intelligence models and approaches for statistical representation are publicly available through a number of websites and links. Refer to these links and access instructions through the following link.
This way we can all work together towards this simple goal.
Photo by National Cancer Institute on Unsplash.
