CHANGES IN VERSION 0.99.31 ------------------------ * ibex_ensure_basilisk_external_dir no longer importsFrom basilisk.utils directly CHANGES IN VERSION 0.99.30 ------------------------ * Moved data processing script out of vignette to inst/scripts * Added ibex_ensure_basilisk_external_dir with basilisk.utils CHANGES IN VERSION 0.99.29 ------------------------ * Adding internal .OnLoad() function to handle basilisk lock dir issue CHANGES IN VERSION 0.99.28 ------------------------ * Deprecated `quietBCRgenes()` * Converted `Ibex.matrix()` to `Ibex_matrix()` * Added Install Instructions for BioCondcutor on README and Vignette * Removed references to Keras3 Installation * Removed LazyData TRUE statement CHANGES IN VERSION 0.99.10 ------------------------ * Added information to example data CHANGES IN VERSION 0.99.9 ------------------------ * Examples now check if python is installed and running CHANGES IN VERSION 0.99.8 ------------------------ * Updated example data to 2k HEL BEAM-Ab from 10x * Converted ibex_example into SCE object for compliance * Large revision of vignette to fit new data/format * Added species argument to runIbex * Updated CoNGA handling of assay for Seurat and Single-Cell Objects. CHANGES IN VERSION 0.99.7 ------------------------ * Integration of Ibex with immApex * Updated Seurat object to v5 * Updated support for SCE format for ```runIbex()``` * Update ```CoNGAfy()``` to function with all versions of Seurat * Updated ```quietBCRgenes()``` to use VariableFeatures() call for SeuratV5 and backward compatibility. * Add ```getHumanIgPseudoGenes()``` to return a list of human Immunoglobulin Pseudo genes that are kept by ```quietBCRgenes()``` ## New Models * Added New Light and Heavy Chain Models * Encoding methods now accepted: "OHE", "atchleyFactors", "crucianiProperties", "kideraFactors", "MSWHIM","tScales", "zScales" * Sequence input: - Human Heavy: 10000000 - Human Light: 5000000 - Human Heavy-Expanded: 5000000 - Human Light-Expanded: 2500000 - Mouse Heavy: 5000000 - Mouse Heavy-Expanded: 5000000 * Trained convolutional and variational autoencoders for Heavy/Light chains - Architecture: 512-256-128-256-512 - Parameters: Batch Size = 128 Latent Dimensions = 128 Epochs = 100 Loss = Mean Squared Error (CNN) & KL Divergence (VAE) Activation = relu Learning rate = 1e-6 - Optimizers: Adam - Early stopping was set to patients of 10 for minimal validation loss and restoration of best weights - CNN autoencoders have batch normalization layers between the dense layers. CHANGES IN VERSION 0.99.6 ------------------------ * Implementing GitHub action workflows * Adding testthat framework * Deprecating clonalCommunity CHANGES IN VERSION 0.99.5 ------------------------ * Added geometric encoding using the BLOSUM62 matrix * Trained classical and variational autoencoders for light/heavy chains with 1.5 million cdr sequences - Architecture: 256-128-30-128-256 - Parameters: Batch Size = 64 Latent Dimensions = 30 Epochs = 100 Loss = Mean Squared Error - Optimizers: Adam - Early stopping was set to patients of 10 for minimal validation loss and restoration of best weights - learn rate varied by models - classical auto encoders have batch normalization layers between the dense layers. CHANGES IN VERSION 0.99.4 ------------------------ * Added chain.checker() function to allow for uncapitlized chain calls CHANGES IN VERSION 0.99.3 ------------------------ * Updated models for manuscript revision - Architecture: 256-128-30-128-256 - Parameters: Batch Size = 64 Learning Rate = 0.001 Latent Dimensions = 30 Epochs = 50 Loss = Mean Squared Error - Optimizers: RAdam (for amino acid properties) and RMSprop (for OHE) - Early stopping was set to patients of 10 for minimal validation loss and restoration of best weights CHANGES IN VERSION 0.99.2 ------------------------ * Updated models to include radam optimization, early stop for min 10 epochs, and all trained on 800,000 unique cdr3s * quietBCRgenes() now does not remove human Ig pseudogenes CHANGES IN VERSION 0.99.1 ------------------------ * Added detection of chain length to function call * Added support for direct output of combineBCR() * Modified quietBCR() to include constant regions and J-chains CHANGES IN VERSION 0.99.0 ------------------------ * Initial commit