Changes in version 2.16.0 o Add support for ARM64 platforms. o Add support for CBCL format in cellCounts. Changes in version 2.12.0 o Improve the data structure used by cellCounts to further improve its speed. o More checks for input parameters to prevent cellCounts from crashing. o Improve the screen output of cellCounts. Changes in version 2.10.0 o Added inbuilt RefSeq annotation for mm39 (mouse genome Build 39). o Streamlined the mapping and counting processes in cellCounts. o Added support for processing dual-index 10x data in cellCounts. Changes in version 2.6.0 o Improved the speed of cellCounts and also reduced its memory use. o Added a parameter 'umi.cutoff' to cellCounts to call all the cells that had a total UMI count greater than the specified threshold. o Added support for FASTQ-format read input in CellCounts. Changes in version 2.4.0 o The 'isPairedEnd' parameter in featureCounts() is now used to check if the type of input reads (paired-end or single-end) is correctly specified. o A new parameter 'countReadPairs' was added to featureCounts to specify if read pairs should be counted for paired-end read data. o Changes to the input parameters and output of cellCounts() function. CellCounts will generate a Sample Sheet for samples included in the scRNA-seq data, based on the sample index set name provided by the user. Structure of the List object returned by cellCounts() is also simplified. cellCounts() now also outputs a BAM file and a gzipped FASTQ file including raw reads for each sample. Changes in version 2.2.0 o Improve cellCounts() on the identification of cell barcodes arising from ambient RNAs. Changes in version 2.0.0 o Rsubread package is ported to Windows OS. o New function cellCounts(): generate UMI counts for Chromium 10X single-cell RNA-seq data. o flattenGTF() function can merge or chop overlap features. o Check and display the amount of memory available on the computer before starting read mapping. o Optimize the data structure used in buildindex() function to reduce its memory use. o qualityScores() function can optionally retrieve quality scores from all the reads. o File paths included in column names of objects returned by featureCounts(), align(), subjunc() and propmapped() functions are removed or shortened where appropriate. o featureCounts() will be terminated if both single-end and paired-end reads are found in the same input file. o Limit on the length of input file names is increased to 1000 bytes for all functions. Changes in version 1.34.0 o New functions: simReads() and scanFasta(). simReads() generates simulation RNA-seq reads for transcripts. o align() and subjunc() estimate fragment length from mapped read pairs and use the estimated length to assist reporting the best alignment. o align() and subjunc() prefer alignments with no indels included over indel-containing alignments when same number of matched bases are found. o align() and subjunc() check if index files were successfully loaded before starting read mapping. o align() and subjunc() detect indels arising from incorrect shifting of seed sequence when being mapped to low-complexity region and exclude such indels from read re-alignment and indel reporting. o buildindex(), align() and subjunc() support gzipped FASTA format. o featureCounts() allows mapped reads to have ‘*’ as their chromosome name. o removeDupReads() supports BAM-format input and output. Changes in version 1.32.0 o New function flattenGTF() that merges overlapping features into a single interval. o New parameter for align() and subjunc(): sortReadsByCoordinates. o New parameters for featureCounts(): readShiftType, readShiftSize and additionalAttributes. o Specify strand protocol for each library individually in featureCounts(). o Much improved speed of align() and subjunc(). o align() and subjunc() return mapping statistics. o Default setting of buildindex() is changed to building a one-block full index.