Statistical methods and software for precision oncology

I am an associate professor in the Department of Biostatistics at the UNC Gillings School of Global Public Health, with a joint appointment at the Lineberger Comprehensive Cancer Center. Our lab develops statistical and machine learning methods for precision oncology, including adaptive clinical trial designs that integrate real-time biomarker data, deep learning methods for tumor subtyping and missing data, and open-source R packages for cancer genomics. We also co-direct the LCCC Biostatistics Shared Resource and lead biostatistics cores for the NCI breast and pancreatic cancer SPOREs.

Research interests: adaptive trial design, Bayesian methods, nonnegative matrix factorization, deep learning, missing data, cancer genomics, precision oncology

Training: PhD, Biostatistics, UNC Chapel Hill (2013). Postdoctoral fellow, Harvard School of Public Health & Dana-Farber Cancer Institute. BS, Biology & Mathematics, Duke University.

Research · CV (PDF) · Publications

Representative Translational Work

Pancreatic oncologists at UNC Lineberger needed a way to classify individual tumors into molecular subtypes from a single biopsy, without requiring a reference cohort. We developed PurIST, a rank-based classifier that handles tumor purity variation, validated it across international cohorts, and worked with the Yeh laboratory to bring it through CLIA certification. It is currently being evaluated prospectively in several clinical trials and has been licensed to Tempus, making it available at hospitals nationwide. That cycle—clinical need, statistical method, validated software, deployed tool—is how most of our projects begin.

PurIST: From Method to Diagnostic

The PurIST (Purity Independent Subtyping of Tumors) classifier, developed in collaboration with the Yeh laboratory, illustrates a complete translational arc:

Its reference-free design enables subtype classification from a single RNA sample without matched normals.

Current Funding

  • MPI, ARPA-H ADAPT program grant (metastatic breast cancer)
  • MPI, NCI U01 (pancreatic cancer)
  • PI, DOD-funded LLM clinical trial navigation tool

See full funding portfolio for details.

Selected Awards

  • 2025 — Gillings Research Excellence Award
  • 2024 — James E. Grizzle Distinguished Alumnus Award
  • 2023 — Teaching Innovation Award, UNC Gillings
  • 2021 — Delta Omega Faculty Award, Gillings School of Global Public Health
  • 2017 — IBM and R.J. Reynolds Junior Faculty Development Award, UNC-CH
  • 2013 — Barry H. Margolin Dissertation Award for best doctoral dissertation

Service & Leadership

  • Breast SPORE Core B Co-Director (P50-CA058223, 2024–2029)
  • Pancreatic SPORE Core C Co-Director (P50-CA257911, 2022–2027)
  • Lineberger LCCC Biostatistics Shared Resource Associate Director (P30-CA016086)
  • Nature Medicine Statistical Advisory Panel (2023–)
  • Associate Editor, Annals of Applied Statistics (2022–)
  • V Foundation Scientific Advisory Board (2023–)
  • TBCRC Statistical Working Group (2017–)
  • Faculty Executive Committee, Department of Biostatistics (2025–)
  • Gillings Research Council (2023–)
  • Chair, Applied Doctoral Exam Committee, Department of Biostatistics (2015–)

Recent Invited Talks

October 2025 Division of Quantitative Sciences, Johns Hopkins Kimmel Cancer Center Baltimore, MD

Bayesian adaptive design and real-time monitoring for metastatic breast cancer platform trials

Presented our ARPA-H ADAPT trial design, including Bayesian borrowing across treatment arms, reinforcement-learning-based allocation, and real-time monitoring tools for a multi-institution metastatic breast cancer platform.

October 2024 James E. Grizzle Distinguished Alumnus Lecture, UNC Gillings Chapel Hill, NC

Replicability, semi-supervised learning and generative AI: recent statistical work in cancer biostatistics

Showed that semi-supervised NMF recovers clinically actionable pancreatic cancer subtypes with higher cross-cohort replicability than standard unsupervised clustering, and presented a generative-AI framework for synthetic clinical trial data that preserves subgroup treatment effects.

June 2024 STATGEN 2024 Conference (Invited Talk) Pittsburgh, PA

Joint Nonnegative Matrix Factorization and Survival Modeling to Select Clinically-relevant Gene Signatures

Introduced a joint NMF-survival objective that selects gene signatures predictive of overall survival in pancreatic cancer, outperforming two-stage approaches on TCGA and ICGC validation cohorts.

News

Lab

Dec 18, 2024

Congratulations to Euphy Wu on successfully defending her PhD in Biostatistics! Euphy’s dissertation, co-mentored by Drs. Naim Rashid and Mike Love, developed methods for allele-specific expression analysis and topic-model-based single-cell clustering.

Site

Aug 15, 2024

The updated Rashid lab website has gone live!

Latest Updates

Selected Publications

  1. Cancer Cell
    The ADAPT learning cancer treatment system: ARPA-H’s initiative to revolutionize cancer therapy
    Andrea H Bild, Michelle C Sangar, Jasmine A McQuerry, Trey Ideker, Scott Kopetz, and 15 more authors
    Cancer Cell, 2026
  2. Stat Med
    Efficient Computation of High‐Dimensional Penalized Piecewise Constant Hazard Random Effects Models
    Hillary M Heiling, Naim U Rashid, Quefeng Li, Xianlu L Peng, Jen Jen Yeh, and 1 more author
    Statistics in Medicine, 2025
  3. JCO
    DNA mutational profiling in patients with colorectal cancer treated with standard of care reveals differences in outcome and racial distribution of mutations
    Federico Innocenti, Wancen Mu, Xueping Qu, Fang-Shu Ou, Omar Kabbarah, and 4 more authors
    Journal of Clinical Oncology, 2024
  4. Biometrics
    Efficient computation of high-dimensional penalized generalized linear mixed models by latent factor modeling of the random effects
    Hillary M Heiling, Naim U Rashid, Quefeng Li, Xianlu L Peng, Jen Jen Yeh, and 1 more author
    Biometrics, 2024
  5. JCGS
    Deeply learned generalized linear models with missing data
    David K Lim, Naim U Rashid, Junier B Oliva, and Joseph G Ibrahim
    Journal of Computational and Graphical Statistics, 2024
  6. JASA
    High-dimensional precision medicine from patient-derived xenografts
    Naim U Rashid, Daniel J Luckett, Jingxiang Chen, Michael T Lawson, Longshaokan Wang, and 6 more authors
    Journal of the American Statistical Association, 2021
  7. CCR
    Purity independent subtyping of tumors (PurIST), a clinically robust, single-sample classifier for tumor subtyping in pancreatic cancer
    Naim U Rashid, Xianlu L Peng, Chong Jin, Richard A Moffitt, Keith E Volmar, and 14 more authors
    Clinical Cancer Research, 2020
  8. JASA
    Modeling between-study heterogeneity for improved replicability in gene signature selection and clinical prediction
    Naim U Rashid, Quefeng Li, Jen Jen Yeh, and Joseph G Ibrahim
    Journal of the American Statistical Association, 2020
  9. Nat Genet
    Virtual microdissection identifies distinct tumor-and stroma-specific subtypes of pancreatic ductal adenocarcinoma
    Richard A Moffitt, Raoud Marayati, Elizabeth L Flate, Keith E Volmar, S Gabriela Herrera Loeza, and 13 more authors
    Nature genetics, 2015