Research Onboarding

Practical starter guide — maps the sidebar sections to GitHub repos and first steps

Welcome — Research Onboarding

The sidebar on the left is the map of the research group's work. Each section matches one kind of project. This page tells you, for every section: what it is, which portal page to open, which GitHub repositories back it, and the smallest sensible first step.

Code root on your laptop: ~/Documents/GitHub · GitHub org: github.com/sbalci

First week: do these in order

Install the core stack

R (≥ 4.3), RStudio, jamovi, QuPath, Python 3.11, Git, VS Code. On macOS also install Xcode command line tools.

Clone the flagship module

git clone https://github.com/sbalci/ClinicoPathJamoviModule. Everything in the jamovi sidebar section descends from this repo.

Walk the sidebar top to bottom

Open each section below in order. Skim the page, note which repos back it, and ask yourself where your assigned project fits.

Reproduce before you change

Once assigned, get the existing analysis building on your machine before editing anything. Reproducing what exists is always the first contribution.

Sidebar sections → what’s behind them

The eight sections below mirror the sidebar in _quarto.yml.

Report Analysis

Parsing and quality-checking of narrative pathology reports — extracting structured fields, running CAP/ICCR style checks, and feeding downstream dashboards.

kris.qmd — Kris reports

GitHub repos

pathology-report-checker-skill — rule-based CAP/ICCR compliance checks
pathology-report-colon-QA — colon-specific QA pilot
patolojirapor, patolojiraporlari — historical report corpora

First step: open pathology-report-checker-skill and run the provided sample report end-to-end.

Text Analysis

Turning free-text reports into structured data. Two layers: a low-level extractor, and a higher-level "report skills" toolkit that wraps LLMs for scoring and rewriting.

extracttext.qmd — Extract Text from Report reportskills.qmd — Skills & Report Quality

GitHub repos

extract-report-text — PDF / HL7 / RTF → plain text pipeline
pathology-reports-text-analysis — NLP over the extracted corpus
pathology-skills-collection, claude-scientific-skills — reusable Claude skills for summarization, translation, checking
n8n-pathology — automation glue
markitdown, repo2file — doc-to-markdown helpers

First step: run the extractor on one de-identified PDF from the anonymized corpus; confirm the JSON output round-trips through the skills collection.

Quality Research

Operational / workflow studies: scanner timings, consultation patterns, turnaround. These power the department's process-improvement papers.

scanningtime.qmd — Scanning Time intradepartmentalconsult.qmd — Intradepartmental Consultation outsideconsults.qmd — Consultations from Outside Labs

GitHub repos

scan-time-in-real-life — scanner utilization & transfer timing
grandQC_quality_study, HistoQC, PathProfiler — slide-level QC inputs
konsultasyon — consultation tracking site
quality-feedback — feedback capture forms

First step: reproduce the hour-of-day figure in scan-time-in-real-life from the shipped sample CSV.

Morphology

Organ-focused morphology studies that combine histology review with structured data capture. Each sub-page owns one organ / question.

pancreasmorphology.qmd — Pancreas Morphology her2gastro.qmd — HER2 Gastro

GitHub repos — pancreas

morphology-research-pancreas-interface, pancreas-adenosquamous, pancreas-solid-pseudopapillary, pdac-large-duct, PDAC-adrenal-invasion, PDAC-benign-vs-malignant, PDAC-cytology, PDAC_LN_shell, ampullary-adenocarcinoma, ectopic-pancreas

GitHub repos — GI / gastric

lymphocytic-gastritis, sarcina-ventriculi, eosinophilic-esophagitis, duodenum-tubular-adenoma, omentum, acute-perforated-appendicitis, extramural-venous-invasion-elastic (TR: ekstramural-venoz-inv-elastik)

Other organ studies available to lift into new sidebar pages

Breast: breast-lobular-carcinoma, HER2-intratumoral-heterogeneity, CSF-breast-carcinoma-metastasis
CNS: glioblastoma, WHO-CNS, secretory-meningioma, arachnoid-cyst, pilocytic-astrocytoma, ganglioneuroma, CSF-metastasis
GU / other: onc_rcc, oncotype, lung-squamous-cell-carcinoma, liver-transplant-acute-rejection, morphea, trypsin, fistula, rokitansky-aschoff-sinus, serous-carcinoma-psammoma-bodies, trps1, gpnmb, parasite

First step: pick one organ repo, reproduce its existing figures, then open a sidebar page for it following the pattern of pancreasmorphology.qmd.

patoloji AI

Applied AI on whole-slide images — both in-house models and vendor integrations (Paige, Aiforia). Each sidebar page is one clinical question or one vendor evaluation.

pinkkidney.qmd — Pink Kidney livertrucutprimary.qmd — Liver Tru-Cut Primary aiforiabreast.qmd — Aiforia Breast paigeprostate.qmd — Paige Prostate

Project repos

PinkKidney, PembeBobrek — kidney AI studies (EN / TR)
morphology-ai-research-liver-interface — liver tru-cut origin AI surface
aiforiabreast — Aiforia breast evaluation
paige-prostate — Paige prostate evaluation

Supporting toolchain

Viewers / WSI I/O: DIYWSI, simple_wsi_viewer, slide_viewer_47, video-wsi, imageTypes
PACS integration: sectra-qupath-extension, sectra-python-qupath-extension
Annotation / curation: QuickAnnotator, HistomicsML, CohortFinder, TissueSelector
Deep-learning pipelines: pathml, pathml-tutorial, tiatoolbox, pathology-hooknet, DL4MicEverywhere
Anonymization / data prep: pathologyAnonymData, pathology_atlas_pipeline

First step: read the three portal SOPs (QuPath annotation, WSI transfer, anonymization) on the home page before touching any patient slide.

Bibliometrics

Analyses of the pathology literature itself — authorship, citation networks, and topic trends.

whociteswho.qmd — WHO Cites Who

GitHub repos

bibliometrics — main analysis notebook
whowriteswho, whowriteswho.github.io — authorship-network site

First step: run the rendering pipeline in bibliometrics against its bundled sample export.

ecosystem

How the group's public-facing properties, internal tools, and repos fit together. Useful when you need to decide where a new piece of work should live.

ecosystem.qmd

Public sites

sbalci.github.io, drozlembalci.github.io, MemorialPath.github.io, patolojiatlasi.github.io, whowriteswho.github.io
research — this portal
patoloji-hakkinda, patoloji-ve-bilisim, rutinde-dijital-patoloji, patoloji-bilgi-yonetim-sistemi, ASDP-podcast, BlueSky

Templates & infrastructure

template, histopathology-template, seed — scaffolds for new projects
shiny-server, intranet, local, data — deployment & local config
MyRCodesForDataAnalysis, histopatoloji-calismalarinda-biyoistatistik-icin-veri-hazirlama, clinicopathological, JamoviTest, SimpleTests, Rj, RwithFileMaker, vtree — reusable R utilities
Archived teaching: EAS1–EAS29, ClinicoPathLinuxDescriptives, MemorialPathStats

First step: clone template once so every new project you start inherits the expected Quarto / R scaffolding.

jamovi

The statistical backbone of every clinicopathological paper in the group. ClinicoPathJamoviModule is the umbrella; the others are focused sub-modules you can install independently.

jamovi.qmd — Overview ClinicoPathDescriptives.qmd jsurvival.qmd meddecide.qmd jjstatsplot.qmd oncopath.qmd — OncoPath

Core modules (one sidebar page each)

ClinicoPathJamoviModule — master module; install first
ClinicoPathDescriptives — Table 1, cross-tabs, stratified descriptives
jsurvival — KM, Cox, competing risks, cure models, ML survival
meddecide — diagnostic test evaluation, decision curves, agreement
OncoPath — oncology-specific templates
jjstatsplot — publication-ready statistical plots

Satellite & specialty modules

ClinicoPathShiny — Shiny web version
ggstatsplot, vijPlots, conttablesplus, jPwrSS, jmvbaseR, gamlj, jamoviPsychoPDA, BrawStats-for-Jamovi-Full
ClinicoPathDescriptives-WIP, meddecide-WIP, jsurvival-WIP — work-in-progress forks; don't ship from these

First step: in jamovi, install the sideload .jmo from ClinicoPathJamoviModule/jamovi/, open the bundled example dataset, and run a Kaplan–Meier analysis. If it renders, your environment is correct.

Conventions that will save you pain

Never commit patient identifiers, raw DICOM/SVS with embedded labels, or un-hashed accession numbers. If unsure, stop and ask before pushing.

One topic, one repo. If a question does not fit the current repo, start a new one from template. Don’t graft unrelated analyses onto an existing project.
Data stays out of git. Large images, PHI, and raw exports live on the Memorial share. Commit only de-identified, analysis-ready tables.
Turkish and English co-exist. Many repos have a Turkish sibling (e.g. extramural-venous-invasion-elastic ↔︎ ekstramural-venoz-inv-elastik, PinkKidney ↔︎ PembeBobrek). Check both before duplicating work.
Use the jamovi module, don’t reimplement it. Consistency across papers beats novelty.
Freeze before you publish. execute.freeze: auto is already set in _quarto.yml; commit the _freeze/ folder so figures reproduce on CI.
Reproduce first, change second. When you inherit a project, get its existing build green on your machine before editing anything.
File naming. Case studies use kebab-case organ-topic names (pancreas-solid-pseudopapillary). Tools use camelCase / PascalCase (ClinicoPathJamoviModule, OncoPath).

Where to ask

PI direction, study approvals: Project Lead.
QuPath, Sectra, scanner issues: Digital Pathology Support.
R, jamovi, statistics: Analysis Support.
This portal or the GitHub organization: open an issue in sbalci/research.

Last updated: 2026-04-20. When you add a new sidebar section, mirror it here so new researchers can see what’s behind it.