Omid Shams Solari
Omid Shams Solari ɔːmɪ̈d ʃæms sɔːlɑːrɪ

Lead AI Research Scientist

University of California, Berkeley
Natera Inc.

About Me

Dr. Omid Shams Solari is currently the Lead AI Research Scientist and Statistician at Natera Inc. His research interests include AI language models, e.g. DNA/protein language models, multi-modal AI, interpretable AI, high-dimensional statistical learning and inference.

Download CV
Interests
  • Explainable AI, Multi-Modal AI, Sequence Models
  • Language Models, e.g. DNA/protein Language Models
  • Statistical/Machine Learning
  • Biomarker Discovery
Education

  • PhD Statistics (Emphasis on Optimization and Machine learning)

    University of California, Berkeley

  • MA Applied Mathematics

    University of California, Berkeley

  • B.Sc. Computational Fluid Dynamics

    Sharif University of Technology

📓 Research Interests
I’m currently a Lead AI Research Scientist and head of Epigenetics Biomarker Discovery Algorithms at Natera Inc. I am currently using language models to predict protein post-translational modifications (PTM) from peptide sequences and model neoantigen immunogenicity as a function of PTM likelihoods in a subsequent vision transformer (ViT) model.
Current Projects
AntigenBinder: An End-To-End Framework for Neo-Antigen Immunogenicity Likelihood Estimation
Cell-Type Decomposition Via Robust Non-Negative Matrix Factorization
XPLoRA: Robust Low-Rank Adaptation of LLMs via Regularized Power Method
BayesBreak
PostMoDisco: A Semi-Supervised Multi-Modal U-Transformer Model for In-Silico Prediction of Post-Translational Modifications
Publications
Omid Shams Solari (2024). AntigenBinder: An End-To-End Framework for Neo-Antigen Immunogenicity Likelihood Estimation.
PDF Code Dataset
Omid Shams Solari (2024). Cell-Type Decomposition Via Robust Non-Negative Matrix Factorization.
Cite Code
Omid Shams Solari (2024). XPLoRA: Robust Low-Rank Adaptation of LLMs via Regularized Power Method.
Cite Code
Omid Shams Solari (2024). BayesBreak. Journal of Source Themes, 1(1).
Cite Code
Omid Shams Solari (2024). PostMoDisco: A Semi-Supervised Multi-Modal U-Transformer Model for In-Silico Prediction of Post-Translational Modifications .
PDF Code Dataset
Preethi Srinivasan, Tzu-Chun Chen, Ehsan Haghshenas, Misagh Kordi, Ilker Tunc, Wenching Chan, Omid Shams Solari, Joshua Babiarz, Yoshiaki Nakamura, Takayuki Yoshino, Others (2024). Circulating differential methylation allele fraction (DMAF) strongly correlates with circulating tumor DNA (ctDNA) variant allele fraction (VAF). Cancer Research.
Cite
Parvathi Myer, Preethi Srinivasan, Sara L Bristow, Shifra Krinshpun, Omid Solari, Vasily N Aushev, Adham a Jurdi, Minetta C Liu, Breeana L Mitchell, Alexey Aleshin, Others (2024). Comparison of colorectal cancer (CRC) characteristics across genetic ancestries: Implications for early cancer detection (ECD).. American Society of Clinical Oncology.
Cite
Omid Solari, Tudor Constantin, Yuqian Jiang, Wengching Chan, Ilker Tunc, Preethi Srinivasan, Misagh Kordi, Marianne Santaguida, Breeana L Mitchell, Alexey Aleshin, Others (2024). Comprehensive analysis of differentially methylated regions in colorectal cancer (CRC).. American Society of Clinical Oncology.
Cite
Parvathi a Myer, Preethi Srinivasan, Sara L Bristow, Shifra Krinshpun, Omid Shams Solari, Vasily N Aushev, Adham Jurdi, Minetta C Liu, Breeana L Mitchell, Alexey Aleshin, Others (2024). Sa1175 COMPARISON OF COLORECTAL CANCER (CRC) CHARACTERISTICS ACROSS GENETIC ANCESTRIES: IMPLICATIONS FOR EARLY CANCER DETECTION. Gastroenterology.
Cite
Omid Shams Solari, Robert Ford (2022). example:. Journal of Source Themes, 1(1).
Cite Code
Omid Shams Solari, Farzad Pourbabaee (2022). Tail Probability Estimation of Factor Models with Regularly-Varying Tails: Asymptotics and Efficient Estimation. Available at SSRN 4078304.
Cite
Erik Gafni, Adam Harvey, Artur Jaroszewicz, Omid Shams Solari, Jane Landolin, Mouadh Barbirou, Amanda Miller, Peter J Tonellato, Anshul Kundaje, Stefanie S Jeffrey, Others (2021). Cell-free DNA fragments inform epigenomic mechanisms for early detection of breast cancer. Cancer Research.
Cite
Omid Shams Solari, Rojin Safavi, James B Brown (2019). BLOCCS: Block Sparse Canonical Correlation Analysis With Application To Interpretable Omics Integration. arXiv e-prints.
Cite
Omid Shams Solari, Rojin Safavi, James B Brown (2019). BLOCCS: Block Sparse Canonical Correlation Analysis With Application To Interpretable Omics Integration. arXiv preprint arXiv:1909.07944.
Cite
Hamutal Arbel, Sumanta Basu, William W Fisher, Ann S Hammonds, Kenneth H Wan, Soo Park, Richard Weiszmann, Benjamin W Booth, Soile v Keranen, Clara Henriquez, Omid Shams Solari, Others (2019). Exploiting regulatory heterogeneity to systematically identify enhancers with high accuracy. Proceedings of the National Academy of Sciences.
Cite
Farzad Pourbabaee, Omid Shams Solari (2019). Large Deviations of Linear Models with Regularly-Varying Tails: Asymptotics and Efficient Estimation. arxiv.
Cite
Omid Shams Solari (2019). Large-Scale Interpretable Multi-View Learning for Very High-Dimensional Problems with.
Cite
Omid Shams Solari (2019). Large-Scale Interpretable Multi-View Learning for Very High-Dimensional Problems with Application to Multi-Omic Data.
Cite
Omid S Solari, James B Brown, Peter J Bickel (2019). Sparse canonical correlation analysis via concave minimization. arXiv preprint arXiv:1909.07947.
Cite
Luisa Orsini, James B Brown, Omid Shams Solari, Dong Li, Shan He, Ram Podicheti, Marcus H Stoiber, Katina I Spanier, Donald Gilbert, Mieke Jansen, Others (2018). Early transcriptional response pathways in Daphnia magna are coordinated in networks of crustacean-specific genes. Molecular Ecology.
Cite
See all publications
Posts

Low-Rank Adaptation

Take full control of your personal brand and privacy by migrating away from the big tech platforms!

MethylSuite

Use popular tools such as Plotly, Mermaid, and data frames.

U-GAN

Use popular tools such as Plotly, Mermaid, and data frames.

Protein Language Models

Easily manage your projects - create ideation mind maps, Gantt charts, todo lists, and more!