v0.2.0 - Multi-trait Support & Usability Improvements

New Features

Multi-trait Phenotype Support

• Analyze multiple traits in a single PANICLE call - pass a phenotype file with ID + multiple trait columns
• Results are organized by trait: results['Trait1']['GLM'], results['Trait2']['MLM'], etc.
• Single-trait usage remains fully backward compatible

Automatic Sample Alignment

• PANICLE now automatically matches sample IDs between phenotype and genotype files
• Samples present in only one file are excluded with clear reporting
• Per-trait exclusion of individuals with missing or non-finite phenotype/covariate values

Convenience Methods

• AssociationResults.manhattan_plot() and .qq_plot() methods for quick visualization directly from results

Improved Data Loading

• Genotype file paths now work directly in PANICLE() (VCF, PLINK .bed, HapMap)
• Loader functions (load_genotype_vcf, load_phenotype_file, etc.) re-exported from top-level panicle namespace
• Case-insensitive and whitespace-tolerant format/trait name matching

Bug Fixes

• Fix unsafe float equality comparisons in LD pruning (BLINK/FarmCPU)
• Suppress spurious floating-point warnings on Apple Silicon (arm64)
• Fix Python 3.9 compatibility (union type syntax)
• Fix mutable default argument in PANICLE() method parameter

Improvements

• Replace print() statements with Python logging module throughout
• Vectorized IBS kinship computation (faster than nested loops)
• Default methods changed to GLM, MLM, FarmCPU (removed BLINK from defaults)
• Better error messages when traits are not found (shows available columns)
• Added CITATION.cff for easy citation via GitHub

Documentation

• Added Python API quickstart with examples in README
• Added CLI usage section with options table
• Fixed docstring/code mismatches