26th, January, 2017 o Development 1.99.8 o Algorithm is now optimized for speed. o Added unit tests. o Compressed all datasets with the smallest possible format. o Added data imputation function. In the manual, we added this function with the note "In development". o Changed parameters in HVar function to make it less stringent. o Allow PCA and HClust to utilize data imputation. Since these two tests are more prone to zero counts, we find that data imputation can help improve their performances. o Note that t-SNE does not need data imputaiton and is the preferred method for cell subpopulation analysis. o PCA is speeded up using gmodels library. o Release 1.2.8 o We now formally declare Linnorm version 1 to be deprecated. o Release 1.2.7 and 1.2.8 are already implemented with many modifications that are intended for Linnorm 2. Still, Linnorm version 1.99.8+ will allow more functionalities. We strongly suggest using the newer version. 17th, January, 2017 o Development 1.99.7, Release 1.2.7 o Finally, build is successful on bioconductor. o minZeroPortoin argument is renamed into minNonZeroPortion (minimum non-zero portion filtering threshold), to better reflect its real function. 2nd, January, 2017 o Development 1.99.0 o Improved Linnorm's algorithm o Added data filtering step o Added normalization step o Added new functions for normalization, t-SNE clustering. o Almost all functions are updated. o Release 1.2.1 o It now matches 1.99.0 as much as possible. 6th, October, 2016 o Development 1.1.16: o Removed warnings from loading package. o Fixed and updated vignettes based on Bioc Staff recommendation. o Implemented Linnorm.HClust function. o Updated Linnorm.Cor function for clustering functionalities. 27th, August, 2016 o Development 1.1.14: o Updated Linnorm.Cor function to allow more functionalities. o Updated vignettes on Linnorm.Cor. o Updated reference manual. o Release: Fixed building error in 1.0.5, thus becoming 1.0.6. 23th, August, 2016 o Critical update for development version 1.1.13: o Added Linnorm.HVar, Linnorm.Cor functions for highly variable gene discovery and gene correlation network analysis. o Added input argument for users who wish to reuse Linnorm transformed datasets in various functions. o Added keepAll argument for users to control whether filtering would remove some genes from the results. o Author list update in release 1.0.5. 17th, August, 2016 o Updated author list. o Error preventing checks now implemented in release 1.0.4. o Critical: Fixed a bug that might cause Linnorm to underestimate fold change. 21st, July, 2016 o Several new updates for the last few versions: o PCA analysis pipeline has been added to Linnorm. o Error preventing checks have been implemented, as suggested by Dr. Jing Qin. Thanks. o Vignetes updated and its graphics are fixed. 16th, March, 2016 o Version 0.99.0 is going to be submitted to Bioconductor Today. 10th, March, 2016 o Congratulations, this is the first news for version 0.99.0.