Overview
This glossary covers terminology commonly used in the Rashid Lab. If you encounter unfamiliar terms, check here first.
Computing & HPC
| Longleaf |
UNC’s high-performance computing (HPC) cluster |
| OnDemand |
Web interface for Longleaf (RStudio, Jupyter, file browser) |
| Slurm |
Job scheduler on Longleaf; manages resource allocation |
| Partition |
Queue for jobs (e.g., general, debug, bigmem) |
| sbatch |
Command to submit a batch job script |
| srun |
Command to run an interactive job |
| Node |
Individual server in the cluster |
| Core/CPU |
Single processing unit; request more for parallel work |
| Job array |
Running many similar jobs efficiently |
R Programming
| data.table |
High-performance data manipulation package (lab standard) |
| targets |
Pipeline toolkit for reproducible workflows |
| crew |
Worker framework for distributed computing with targets |
| DESCRIPTION |
File listing package dependencies |
| .Rprofile |
User configuration file that runs when R starts |
| here::here() |
Function to create project-relative paths |
| roxygen2 |
Documentation system for R packages |
Git & GitHub
| Repository (repo) |
Project folder tracked by Git |
| Commit |
Snapshot of code at a point in time |
| Branch |
Parallel version of code for feature development |
| Pull Request (PR) |
Proposal to merge changes; requires review |
| Issue |
Tracked task, bug, or discussion |
| Discussion |
Forum-style conversation (we use for meeting agendas) |
| Fork |
Personal copy of someone else’s repository |
| Clone |
Download a repository to your local machine |
| Push/Pull |
Send/receive commits to/from remote repository |
| Merge |
Combine branches together |
| Squash |
Combine multiple commits into one |
Statistical Methods
| Bayesian Optimization (BO) |
Method for efficiently optimizing expensive functions |
| Gaussian Process (GP) |
Probabilistic model; provides uncertainty estimates |
| Acquisition function |
Guides where to sample next (e.g., Expected Improvement) |
| Adaptive trial |
Clinical trial that modifies based on interim data |
| Operating characteristics |
Performance metrics (power, type I error, sample size) |
| Calibration |
Finding design parameters that meet statistical constraints |
| Multi-fidelity |
Using low/high accuracy simulations strategically |
| Warmstart |
Beginning optimization with prior information |
| Posterior probability |
Updated belief after seeing data |
| Prior |
Initial belief before seeing data |
| Futility |
Stopping early because treatment effect unlikely |
| Efficacy |
Stopping early because treatment clearly works |
Clinical Trials
| Platform trial |
Multi-arm trial that adds/drops treatments over time |
| Single-arm (SA) |
Trial comparing treatment to historical control |
| Between-arm (BA) |
Trial with concurrent control group |
| Hybrid SA→BA |
Trial starting single-arm, adding control if promising |
| Type I error |
False positive rate (rejecting null when true) |
| Power |
True positive rate (rejecting null when false) |
| Expected sample size (EN) |
Average patients enrolled under a scenario |
| N_max |
Maximum sample size cap |
| Interim analysis |
Looking at data before trial ends |
| Stopping boundary |
Threshold for early termination |
| IRB |
Institutional Review Board (ethics approval) |
| DSMB |
Data Safety Monitoring Board |
Lab Workflow
| Meeting lead |
Person facilitating that week’s lab meeting |
| 1:1 |
One-on-one meeting with Dr. Rashid |
| Action item |
Task assigned during a meeting |
| Parking lot |
Items deferred to discuss later |
| Discussion thread |
GitHub Discussions post for meeting agenda/notes |
| Project handoff |
Transferring a project to another lab member |
Reproducibility
| Seed |
Number ensuring random processes are reproducible |
| Pipeline |
Automated sequence of analysis steps |
| Provenance |
Record of where data came from and how it was processed |
| Config file |
Central place for project parameters (e.g., globals.yml) |
| YAML |
Human-readable configuration format |
| Symlink |
Shortcut linking to a file/folder elsewhere |
UNC Resources
| Onyen |
UNC username |
| ConnectCarolina |
UNC administrative portal |
| ITS |
Information Technology Services |
| Research Computing |
Team managing Longleaf and HPC resources |
| Gillings |
UNC Gillings School of Global Public Health |
| BIOS |
Department of Biostatistics |
| LCCC |
Lineberger Comprehensive Cancer Center |
Acronyms
| HPC |
High-Performance Computing |
| BO |
Bayesian Optimization |
| GP |
Gaussian Process |
| EI |
Expected Improvement (acquisition function) |
| LHS |
Latin Hypercube Sampling |
| CI |
Confidence Interval |
| EN |
Expected sample size under null/alternative |
| SA |
Single-Arm |
| BA |
Between-Arm |
| PR |
Pull Request |
| PHI |
Protected Health Information |
| HIPAA |
Health Insurance Portability and Accountability Act |
| IRB |
Institutional Review Board |
| CITI |
Collaborative Institutional Training Initiative |
| JASA |
Journal of the American Statistical Association |
| Biometrics |
Biostatistics journal |
Adding Terms
If you encounter a term that should be here:
- Edit this file and add the term
- Submit a PR or mention it in lab meeting
- Help the next person who joins!
See also: Tools Setup | R Style Guide | Git Practices