feat: Add interactive Shiny application, startpage landing, and dockerized deployment

[JD2112]

@lovisaor

Summary of Changes

Key Features & Modules

1. Interactive Workflow Tabs:

   • Start Page & Pipeline Routing (module_startpage.R): Dedicated landing page with digital biology hero visuals, researcher profile links (LiU BKV / ÖstLab), publication DOI, and dynamic routing between Example Data and Custom Data pipelines.
   • Load / Create PAC (module_load.R): Seamlessly load the built-in Drosophila PAC dataset (9,131 seqs x 9 samples), upload .rds/.RData PAC objects, or generate new PAC objects from raw FASTQs + Pheno CSV with trimming controls.
   • Filter & Normalize (module_filter.R): Interactive sliders and selectors for size filtering (18–35 nt), minimum counts, coverage thresholds, and CPM/VST/RLOG normalizations with reactive summary metrics.
   • Annotation Explorer (module_annotate.R): Searchable, interactive DT tables and category breakdown charts for multi-mapping annotation hierarchies.
   • Post-Filtering Analysis (module_analyze.R): Comprehensive analysis suite including:
     • 2D/3D PCA visualization
     • Interactive DESeq2 differential expression volcano plots and tables
     • Size distribution and 5' nucleotide bias histograms
     • Composition stacked bars and pie charts

2. UX & Design System (inst/shiny/www/style.css):

   • Modern, high-contrast solid design system with Google Fonts (Plus Jakarta Sans, Outfit, and JetBrains Mono).
   • Standardized 100% zoom web proportions with zero edge gaps on the header and 3-column footer.
   • Slide-in offcanvas drawer for live execution logs (.side-log-drawer).
   • Floating footer navigation controls for switching analysis modes.

3. Package & Deployment Integration:

   • Added run_seqpac_app() in R/run_app.R to launch the Shiny application directly from an interactive R session.
   • Added Dockerfile (rocker/shiny-verse:latest) containerizing all CRAN/Bioconductor dependencies to serve the app on port 3838.
   • Added setup and verification guide in run-shiny.md.

Bug Fixes & Stability

• S4 PAC Object Deserialization: Added to_s4_pac() helper to properly handle serialized S4 PAC objects from uploaded .rds files without encountering $ operator not defined for this S4 class.
• DESeq2 Column Parsing: Corrected log2FC regex pattern matching for accurate up/down-regulated sequence counts.

Testing & Verification

• All R scripts in inst/shiny/ have been verified for syntax with parse().
• Successfully built and tested against local Docker container (rocker/shiny-verse:latest).
• Verified both built-in example dataset analysis and custom PAC .rds file uploads.

Checklist

• Code follows existing package naming and formatting conventions
• Documentation & CHANGELOG.md updated
• Zero syntax errors (R CMD check / script validation passed)
• Tested locally and via Docker container