v1.1

What's New in v1.1

• Free to use — removed all license key requirements and trial restrictions. No activation needed.

v1.0.4

Fixes

Streamlit not found on first launch

macOS .app bundles use a minimal shell environment, causing nohup to lose the conda PATH.

• Use explicit paths to streamlit and pip binaries instead of relying on conda activate
• Verify streamlit was actually installed before marking setup as complete

Child processes (Salmon, FastQC, Trim Galore) left running after quit

The launcher used nohup + exit, leaving Streamlit and pipeline tools as orphaned processes.

• Launcher now uses setsid + process group cleanup via trap to kill all children on exit

Custom index build fails / takes 30+ minutes

Building from a user-provided FASTA was missing the --gencode flag and tx2gene generation.

• Auto-detect GENCODE format and add --gencode flag when appropriate
• Generate tx2gene mapping for GENCODE FASTAs (matches built-in reference behavior)

DESeq2 not found (PCA and DE analysis fail)

The single large Bioconductor conda install could silently fail due to dependency conflicts.

• Split into smaller install batches for reliable dependency resolution
• Add missing bioconductor-biomart package
• Verify DESeq2 is actually loadable before marking setup as complete

If upgrading

Delete ~/.rnaseq-processor and reinstall from this DMG to trigger a fresh setup.

v1.0.3

Fix: Too many open files crash on macOS

Fixed OSError: [Errno 24] Too many open files that occurred when launching the app from Finder.

Root cause: macOS app bundles launched via Finder inherit a 256 file descriptor soft limit from launchd, which is too low for Streamlit + the scientific Python stack (pandas, plotly, numpy, etc.).

Fixes:

• Raise file descriptor limit to 8192 in Python via resource.setrlimit before heavy imports
• Raise shell-level limit via ulimit -n 4096 in the launcher script
• Disable Streamlit file watcher (--server.fileWatcherType none) since hot-reload is unnecessary in a bundled app

v1.0.2 - Metadata editor fix & UI cleanup

Changes

• Fixed metadata editor glitch where edits were lost when navigating between cells
• Added form-based editing with explicit "Save Metadata" button
• Removed session save/load feature from sidebar
• New app icon (Genomify with Garrett logo)
• Improved DMG installer with drag-to-Applications layout

Previous (v1.0.1)

• Added optional FASTQC and Trim Galore preprocessing steps
• Custom run naming for output folders
• Improved volcano plot geneset overlay with static R-generated export

v1.0.1 - UI improvements and preprocessing

Changes

• Added optional FASTQC and Trim Galore preprocessing steps
• Custom run naming for output folders
• Improved volcano plot geneset overlay with static R-generated export (PNG, PDF, TIFF)
• Fixed time estimates based on selected preprocessing options
• Cleaner UI - removed kaleido dependency messages
• Fixed run name text field persistence
• Smaller circle outlines on volcano plot gene set highlights

RNA-seq Processor v1.0.0

Installation

1. Download RNA-seq-Processor-Installer-1.0.0.dmg
2. Open the DMG
3. Drag RNA-seq Processor to Applications
4. Launch from Applications
5. On first launch, click Install to set up dependencies (~2GB download, ~15 min)

Your Data Location

All data is saved to ~/Documents/RNAseq_Processor/:

• indices/ - Downloaded reference genomes
• results/ - Quantification output (includes QC reports)
• sessions/ - Saved analysis sessions

Features

• Preprocessing pipeline
  • FASTQC quality reports (automatic)
  • Trim Galore adapter trimming (automatic)
• Salmon quantification for RNA-seq data
• Differential expression analysis (DESeq2)
• Gene set enrichment analysis (fgsea)
• Interactive volcano plots
• PCA visualization
• TIFF/PNG export for all figures
• Session save/load
• Real-time progress tracking

Requirements

• macOS 10.15 or later (Intel or Apple Silicon)
• ~3GB disk space for dependencies
• Internet connection for first-time setup