NEWS

---

SeqArray 1.53.4

NEW FEATURES

• new argument '.package' in 'seqParallel()', 'seqApply()', 'seqBlockApply()' and 'seqUnitApply()' for the packages loaded before calling the user-defined function, so that the functions in these packages can be used directly: they are loaded in each child process of a cluster or a 'BiocParallelParam' object, or in the current R process before forking or when only one process is used

UTILITIES

• 'seqGetData(, "position")', "$chrom_pos", "$chrom_pos2" and "$chrom_pos_allele", and 'seqApply()' with these variables no longer load the positions of all variants into memory: the positions are read from the GDS file via a cache of up to 131072 consecutive variants per file, which is efficient when the variants are accessed in order (e.g., 'seqBlockApply()' and 'seqUnitApply()')
• 'seqSetFilterChrom(, from.bp=, to.bp=)', and so 'seqSetFilter()' with GRanges or IRanges and 'seqSetFilterPos()', no longer load the positions of all variants into memory: only the minimum and maximum positions of every 131072 variants on a chromosome are kept, the chunks of variants out of the requested ranges are skipped without reading, and binary search is used in a chunk if its positions are sorted
• 'seqFilterPush()' and 'seqSetFilter(, action="push")', also used internally by 'seqBlockApply()', 'seqUnitApply()', etc., store the pushed sample and variant filters as bit vectors instead of byte vectors, so each pushed filter uses 1/8 of the memory (1 bit per sample or variant) until it is popped back
• to reduce memory usage, 'seqOpen()' no longer allocates the sample and variant filters (1 byte per sample or variant), since all samples and variants are selected; a filter is allocated only when it is needed (e.g., to set a filter or read data), and resetting the filter (e.g., 'seqResetFilter()') does not release the allocated memory, which is released when the file is closed
• the internal function '.buffer_position()' is removed, since the positions of all variants are no longer loaded into memory

BUG FIXES

• 'seqGetData(, "position")' and 'seqSetFilterChrom(, from.bp=, to.bp=)' used the old positions after 'seqAddValue(, varnm="position")' replaced them in the opened file
• 'seqSetFilterAnnotID()' ignores "" and "." in 'id', instead of selecting all variants without an ID
• 'seqVCF2GDS()' failed with "more columns than what expected" on a VCF file with a FORMAT column but no sample

SeqArray 1.53.3

NEW FEATURES

• AARCH64 NEON SIMD optimization for vectorized internal functions, providing hardware-accelerated performance on Apple Silicon and other ARM64 platforms
• 'seqParallel()' supports a 'BiocParallelParam' object for the 'cl' argument, with full feature parity to the cluster-based path, including load balancing, init/finalize hooks, and numeric/NULL 'gdsfile' inputs

UTILITIES

• 'seqVCF2GDS(, parallel=n)': each parallel job seeks to its starting position directly instead of reading the file from the beginning (for a plain text or BGZF file, but not a gzip file), no tabix index is needed, and the temporary files are merged while the other jobs are still running
• new ‘split' in 'seqVCF2GDS()': "block" (default) splits BGZF files at the block boundaries without counting the variants; "variant" counts the variants (in parallel for BGZF files) and splits by the variant index, which is used if not all files are BGZF or ’start', 'count' or 'variant_count' is given
• SeqArray no longer needs the Rsamtools package: reading and writing the BGZF format (i.e., bgzip) and building the .csi index are implemented in the package using zlib, so they work on all platforms including Windows; 'use_Rsamtools' in 'seqVCF_Header()', 'seqVCF2GDS()' and 'seqGDS2VCF()' is deprecated and ignored
• 'seqGDS2VCF()' always outputs a BGZF file (i.e., bgzip) when the file name ends with .gz or .bgz, together with a .csi index instead of a .tbi index, since the position in a .tbi index is limited to 2^29 (~512Mb)
• ‘seqUnitMerge()' now properly supports merging more than two ’SeqUnitListClass' objects via '...'
• new 'seqUnitSetDiff()' to remove variant indices of one unit list from another, returning the set-difference version of units
• 'val' in ‘seqAddValue()' can be a function returning the data block by block, when adding a variable to "annotation/info": the blocks are written to the GDS file one by one, so the memory usage does not depend on the total number of variants (’packed' and 'packed.idx' are ignored in this case, since the data are unknown before writing). New 'param' in ‘seqAddValue()', passed to the second argument of the user-defined function ’val(k, param)'

BUG FIXES

• ‘seqVCF2GDS(, start=, count=, parallel=n)' returned more variants than requested, and split the variants wrongly when ’start' was not 1; 'variant_count' should be exact now, since the variants beyond it are not converted
• ‘seqSetFilterPos()' errored for an empty ’pos' vector (e.g., ‘integer(0)'), or when ’chr' contained ‘NA' values (e.g., 'c("1", NA, "2")'); an empty ’pos' now clears the variant filter, and 'NA' chromosomes simply do not match, with 'ret.idx' staying aligned with the query
• ‘seqGetData()' and 'seqBlockApply()' failed with "LENGTH or similar applied to S4 object" when reading a FORMAT variable stored in a sparse array (e.g., "sp.real32" or "sp.real8u"); a sparse matrix is returned with the dimnames ’sample' and 'variant', or a list of sparse matrices if '.tolist=TRUE'
• 'seqGetData(, .tolist=TRUE or NA)' and 'seqBlockApply(, .tolist=TRUE or NA)' on a FORMAT variable: each matrix in the returned list should have the number of selected samples, instead of the total number of samples, when there is a sample filter

SeqArray 1.52.1

NEW FEATURES

• 'seqParallel()' can load cloud-related packages on worker processes before calling the user-defined function, with backward compatibility for older gdsfmt versions

UTILITIES

• faster 'seqSetFilterPos()' than v1.52.0
• 'seqApply()' supports "$ref", "$alt", "$chrom_pos", "$chrom_pos_allele" and "$variant_index" as variable names, consistent with 'seqGetData()'
• 'seqAddValue()' supports 'LogicalList', 'IntegerList', 'NumericList' and 'CharacterList' from the IRanges package as variable-length annotation data input
• 'seqMulticoreSetup()' invisibly returns the newly created cluster object when a multicore cluster is set up

SeqArray 1.52.0

NEW FEATURES

• new 'header' in 'seqGDS2VCF()' to specify whether exporting the metadata header to a VCF file or not
• set 'nosample=TRUE' in 'seqGDS2VCF()' to exclude the sample-level data
• new 'verbose.progress' in 'seqGDS2VCF()' to control the progress display
• 'seqGet2bGeno()' allows a GDS file name in the first argument
• new 'parallel' in 'seqGet2bGeno()' for parallel loading
• new 'SeqArray:::process_block_index' and 'SeqArray:::process_block_count' used in children processes when balancing work
• new 'ploidy' in 'seqVCF2GDS()' and 'seqBCF2GDS()'

UTILITIES

• Reduce memory usage in 'seqParallel()' by avoiding the transfer of unused data during work balancing in parallel
• minor fix in 'seqAsVCF()' when loading VariantAnnotation
• fix the package anchors for all Rd \link{}
• Updated 'seqSetFilterPos()' is significantly faster than the old version

BUG FIXES

• fix 'seqApply(..., margin="by.sample")' dimension mismatch error when multiallelic variants are present

SeqArray 1.50.1

UTILITIES

• Reduce memory usage in 'seqMerge()' when there are many INFO variables

SeqArray 1.50.0

NEW FEATURES

• new '.status_file' & '.proc_time' in 'seqParallel()' for controlling the display of the status of child processes
• enhance workload balancing within 'seqParallel()'
• new 'balancing=NA' in 'seqAlleleFreq()', 'seqAlleleCount()', 'seqGetAF_AC_Missing()', 'seqSetFilterCond()', 'seqUnitFilterCond()' (using workload balancing by default)
• new '.balancing' in 'seqApply()' and 'seqBlockApply()'
• new 'alt' & 'ns' in 'seqGetAF_AC_Missing()'
• new 'dosage' in 'seqExampleFileName()'

UTILITIES

• 'seqDigest()' allows a GDS file name in the first argument
• a new option 'parallel' in 'seqDigest()'
• ‘seqVCF2GDS()' and 'seqBED2GDS()': save the gds node ’chromosome' before 'position'
• ‘seqAddValue(, varnm="annotation/filter", desp=...)' adds ’Description' via the argument 'desp'
• tweak the display of 'seqGDS2BED()'
• 'seqBlockApply()' and 'seqApply()' allows a GDS file name in the first argument
• 'seqGDS2VCF()' generates an indexing file (.tbi) according to the vcf.gz output file, requiring the Rsamtools package
• the first argument of 'seqOptimize()' can be a SeqVarGDSClass object

BUG FIXES

• Previous ‘seqGetAF_AC_Missing(, minor=FALSE)' returns the AF for the alt. alleles and the AC for the ref. alleles, when a dosage GDS file is used; a new ’alt' is used to specify the ref. or alt. alleles.
• Fix '##INFO=<>' in the output of 'seqGDS2VCF()' when the INFO variable is added by users

SeqArray 1.48.0

NEW FEATURES

• ‘seqAddValue()': use bit1 for a logical vector; new argument ’use_float32=FALSE' for storing double
• new argument 'start' in 'seqResetVariantID()'
• new argument 'digest' in 'seqRecompress()' to add MD5 hash codes
• ‘seqGetData(, "$chromosome")' returns chromosome codes in an object of ’S4Vectors::Rle'
• 'seqGetData(, .tolist=NA)' returns an extended list defined in IRanges (e.g., IntegerList) when it is applicable
• 'seqListVarData(, useList=TRUE)' returns an extended list defined in IRanges

UTILITIES

• Tweak display in 'seqResetVariantID()'
• use 'crayon::silver()' instead of 'crayon::blurred()' in the display since RStudio blurs the screen output

BUG FIXES

• 'seqBlockApply()' should recover the filter when the user-specified function fails

SeqArray 1.46.2

UTILITIES

• ‘seqVCF_Header()' allows multiple cores to calculate the total number of variants when ’getnum=TRUE' (the Rsamtools package should be installed); 'seqVCF2GDS()' is faster when obtaining the number of variants for splitting files.
• new 'variant_count' in ‘seqVCF2GDS()' to specify the number of variants in the VCF file when it is known or an approximation is known; it is only applicable when multiple cores are used. If ’variant_count' is specified, counting the number of variants will be skipped.

SeqArray 1.46.0

UTILITIES

• ‘seqGetData()' return NULL, if ’var.name=character()'

SeqArray 1.44.3

UTILITIES

• update the C codes according to '_R_USE_STRICT_R_HEADERS_=true' & '_R_CXX_USE_NO_REMAP_=true'

SeqArray 1.44.2

BUG FIXES

• fix 'seqAddValue(, val=vector("list", NUM_VARIANT))'
• fix the ploidy returned from 'seqVCF_Header()', when there are genotypes of males and females on Chromosome X

SeqArray 1.44.1

UTILITIES

• new option 'numvariant' in 'seqEmptyFile()'

BUG FIXES

• 'seqMerge()' should internally use "chr_position_ref_alt" to distinguish the variants in different files
• 'seqAddValue(, varnm="annotation/filter")' should work with a factor variable
• 'seqAddValue(, varnm="variant.id")' can reset the variant IDs with a different number of the variants

SeqArray 1.44.0

UTILITIES

• tweak the display of progress information in 'seqVCF2GDS()'
• 'seqVCF_Header(, getnum=TRUE, verbose=TRUE)' to show the progress information for scanning the VCF file
• new 'seqGetData(, "$dosage_alt2")' and 'seqGetData(, "$dosage_sp2")' for sex chromosomes, when the alleles are partially missing (e.g., genotypes on chromosome X for males)
• new 'verbose.clean' in 'seqExport()' to control how much information to be displayed

SeqArray 1.42.4

BUG FIXES

• 'seqGetData(, "$dosage_alt")' and 'seqGetData(, "$dosage_sp")' work correctly when the ploidy is >2 and there are missing alleles
• fix a bug that ‘seqParallel()' does not call a user-defined ’.combine' when 'parallel=1'

SeqArray 1.42.1

UTILITIES

• update the help files of 'seqBlockApply()' and 'seqUnitApply()'
• detect the output filename extension in 'seqGDS2VCF()' without considering the case of the characters, supporting .gz, .bgz, .bz and .xz as a filename extension
• fix the compiler warning: -Wformat-security
• new option 'include.pheno=TRUE' in 'seqBED2GDS()'

SeqArray 1.42.0

UTILITIES

• new option 'write.rsid' in 'seqGDS2BED()'

SeqArray 1.40.1

BUG FIXES

• 'seqAddValue(gdsfile, varnm="position")' works correctly

SeqArray 1.40.0

• fix the compiler warning: sprintf is deprecated

SeqArray 1.38.0

UTILITIES

• new option 'ext_nbyte' in 'seqGet2bGeno()'
• 'seqAlleleCount()' and 'seqGetAF_AC_Missing()' return NA instead of zero when all genotypes are missing at a site
• 'seqGDS2VCF()' does not output the FORMAT column if there is no selected sample (e.g., site-only VCF files)
• 'seqGetData(, "$chrom_pos2")' is similar to 'seqGetData(, "$chrom_pos")' except the duplicates with the suffix ("_1", "_2" or >2)

NEW FEATURES

• 'seqGDS2BED()' can convert to PLINK BED files with the best-guess genotypes when there are only numeric dosages in the GDS file
• 'seqEmptyFile()' outputs an empty GDS file

SeqArray 1.36.2

BUG FIXES

• fix the bug at multi-allelic sites with more than 15 different alleles, see https://github.com/zhengxwen/SeqArray/issues/78

SeqArray 1.36.1

BUG FIXES

• 'seqExport()' failed when there is no variant
• 'seqSetFilter(, ret.idx=TRUE)', see https://github.com/zhengxwen/SeqArray/issues/80

SeqArray 1.36.0

NEW FEATURES

• new functions 'seqUnitCreate()', 'seqUnitSubset()' and 'seqUnitMerge()'
• new functions 'seqFilterPush()' and 'seqFilterPop()'
• new functions 'seqGet2bGeno()' and 'seqGetAF_AC_Missing()'
• new function 'seqGetData(, "$dosage_sp")' for a sparse matrix of dosages
• the first argument 'gdsfile' can be a file name in 'seqAlleleFreq()', 'seqAlleleCount()', 'seqMissing()'
• new function ‘seqMulticoreSetup()' for setting a multicore cluster according to a numeric value assigned to the argument ’parallel'

UTILITIES

• allow opening a duplicated GDS file ('allow.duplicate=TRUE') when the input is a file name instead of a GDS object in 'seqGDS2VCF()', 'seqGDS2SNP()', 'seqGDS2BED()', 'seqVCF2GDS()', 'seqSummary()', 'seqCheck()' and 'seqMerge()'
• remove the deprecated '.progress' in 'seqMissing()', 'seqAlleleCount()' and 'seqAlleleFreq()'
• add 'summary.SeqUnitListClass()'
• no genotype and phase data nodes from 'seqSNP2GDS()' if SNP dosage GDS is the input

BUG FIXES

• ‘seqUnitApply()' works correctly with selected samples if ’parallel' is a non-fork cluster
• 'seqVCF2GDS()' and 'seqVCF_Header()' work correctly if the VCF header has white space
• 'seqGDS2BED()' with selected samples for sex and phenotype information
• ‘seqGDS2VCF()' failed if there is no ’genotype/data' in the GDS file

SeqArray 1.32.0

NEW FEATURES

• new option 'ret.idx' in 'seqSetFilter()' for unsorted sample and variant indices
• new option 'ret.idx' in 'seqSetFilterAnnotID()' for unsorted variant index
• rewrite the function ‘seqSetFilterPos()': new options ’ref' and 'alt', 'multi.pos=TRUE' by default
• new option 'packed.idx' in 'seqAddValue()' for packing an indexing variable
• new option 'warn' in 'seqSetFilter()' to enable or disable the warning
• new functions 'seqNewVarData()' and 'seqListVarData()' for variable-length data

UTILITIES

• allow no variant in 'seqApply()' and 'seqBlockApply()'
• the list object returned from 'seqGetData()' always have names if there are more than one input variable names

BUG FIXES

• 'seqGDS2VCF()' should output "." instead of NA in the FILTER column
• ‘seqGetData()' should support factor when ’.padNA=TRUE' or '.tolist=TRUE'
• fix 'seqGDS2VCF()' with factor variables
• ‘seqSummary(gds, "$filter")' should return a data frame with zero row if ’annotation/filter' is not a factor

SeqArray 1.30.0

UTILITIES

• show a warning when an unsorted index is used in 'seqSetFilter()'
• show a message if 'seqVCF_Header()' fails
• a new option 'chr_prefix' in 'seqGDS2VCF()'

BUG FIXES

• ‘seqVCF_Header()' fixes ’contig' in the header of VCF if there are different fields

SeqArray 1.28.1

BUG FIXES

• 'seqRecompress(, verbose=FALSE)' works correctly
• 'seqSetFilter(, action="push+set")' should not reset the filter before setting a new filter

SeqArray 1.28.0

NEW FEATURES

• new functions 'seqUnitSlidingWindows()', 'seqUnitApply()', 'seqUnitFilterCond()'
• new variable "$variant_index", "$sample_index" in 'seqGetData()', 'seqBlockApply()' and 'seqUnitApply()' to get the indices of selected variants
• new arguments '.padNA' and '.envir' in 'seqGetData()'
• new functions 'seqSetFilterAnnotID()' and 'seqGDS2BED()'
• multicore function in 'seqBED2GDS(, parallel=)'
• new package-wide option 'options(seqarray.nofork=TRUE)' to disable forking
• new option 'minor' in 'seqAlleleFreq()' and 'seqAlleleCount()'
• new option 'verbose' in ‘seqMissing()', 'seqAlleleFreq()' and 'seqAlleleCount()'; ’.progress' is deprecated, but still can be used for compatibility
• ‘seqAlleleFreq()', 'seqAlleleCount()', 'seqMissing()', 'seqSetFilterCond()' work on ’annotation/format/DS', if 'genotype/data' is not available

UTILITIES

• 'seqAddValue()' adds vectors, matrices and data frame to "annotation/info"
• 'seqBED2GDS()' allows a single file name without the extended file names (.bed, .fam, .bim)
• allele flip in 'seqBED2GDS()' to allow major allele to be reference
• rewrite 'seqGetData()' for faster loading
• significantly improve ‘seqBlockApply()' on ’annotation/info/VARIABLE' (https://github.com/zhengxwen/SeqArray/issues/59)
• add a S3 method 'print.SeqVCFHeaderClass()' for 'seqVCF_Header()'
• new option '.tolist' in 'seqGetData()', 'seqBlockApply()' and 'seqUnitApply()'
• String "." in a VCF file are converted to a blank string (missing value) in 'seqVCF2GDS()'
• add a class name 'SeqVarDataList' to the returned 'list(length, data)' from 'seqGetData()'
• new option 'seqMissing(, per.variant=NA)'
• add 'comment.char=""' to 'seqBED2GDS()'

SeqArray 1.26.2

NEW FEATURES

• multiple variable names are allowed in 'seqGetData(, var.name=)'

BUG FIXES

• fix 'seqGetData(, "genotype", .useraw=NA)' (https://github.com/zhengxwen/SeqArray/issues/58)

SeqArray 1.26.1

BUG FIXES

• fails to correctly select duplicate indices in 'seqSetFilter(f, variant.sel=)'

SeqArray 1.26.0

NEW FEATURES

• new function 'seqAddValue()'

UTILITIES

• RLE chromosome coding in 'seqBED2GDS()'
• change the file name "vignettes/R_Integration.Rmd" to "vignettes/SeqArray.Rmd", so 'vignette("SeqArray")' can work directly
• correct Estimated remaining Time to Complete (ETC) for load balancing in 'seqParallel()'

BUG FIXES

• 'seqBED2GDS(, verbose=FALSE)' should have no display

CHANGES

• use a svg file instead of png in vignettes

SeqArray 1.24.2

NEW FEATURES

• add the compiler information in 'seqSystem()'
• new arguments '.balancing', '.bl_size' and '.bl_progress' in 'seqParallel()' for load balancing

UTILITIES

• improve unix forking processes for load balancing in 'seqParallel()'

BUG FIXES

• fix 'seqSummary()' when no phase data

SeqArray 1.24.0

NEW FEATURES

• a new function 'seqResetVariantID()'
• a new option in 'seqRecompress(, compress="none")' to uncompress all data
• 'seqGetData()' allows a GDS file name in the first argument

SeqArray 1.22.6

BUG FIXES

• 'seqSetFilter(, sample.id=)' fails to correctly select samples in a few cases (since SeqArray>=v1.22.0 uses the distribution of selected samples to optimize the data access of genotypes, see https://github.com/zhengxwen/SeqArray/issues/48)
• the bgzf VCF file is truncated in 'seqGDS2VCF()' since the file is not closed appropriately
• invalid chromosomes and position in the output of 'seqMerge()' when merging different samples but same variants

SeqArray 1.22.3

NEW FEATURES

• a new option 'scenario' in 'seqVCF2GDS()' and 'seqBCF2GDS()'

UTILITIES

• more information in 'seqDelete()'

BUG FIXES

• export a haploid VCF file using 'seqGDS2VCF()'
• export VCF without any FORMAT data in 'seqGDS2VCF()'
• export GDS without genotypes in 'seqExport()'
• fix parallel file writing in seqVCF2GDS(), when no genotype

SeqArray 1.22.0

NEW FEATURES

• 'seqSNP2GDS()' imports dosage GDS files
• 'seqVCF_Header()' allows a BCF file as an input
• a new function 'seqRecompress()'
• a new function 'seqCheck()' for checking the data integrity of a SeqArray GDS file
• 'seqGDS2SNP()' exports dosage GDS files

UTILITIES

• avoid duplicated meta-information lines in 'seqVCF2GDS()' and 'seqVCF_Header()'
• require >= R_v3.5.0, since reading from connections in text mode is buffered
• 'seqDigest()' requires the digest package
• optimization in reading genotypes from a subset of samples (according to gdsfmt_1.17.5)

BUG FIXES

• 'seqVCF2GDS()' and 'seqVCF_Header()' are able to import site-only VCF files (i.e., VCF with no sample)
• fix 'seqVCF2GDS()' and 'seqBCF2GDS()' since reading from connections in text mode is buffered in R >= v3.5.0

SeqArray 1.20.1

BUG FIXES

• 'seqExport()' fails to export haploid data (e.g., Y chromosome)
• 'seqVCF2GDS()' fails to convert INFO variables when Number="R"

SeqArray 1.20.0

NEW FEATURES

• 'seqGDS2VCF()' outputs a bgzip vcf file for tabix indexing
• two more options "Ultra" and "UltraMax" in 'seqStorageOption()'
• '@chrom_rle_val' and '@chrom_rle_len' are added to a GDS file for faster chromosome indexing
• new function 'seqBCF2GDS()' (requiring the software bcftools)
• new function 'seqSetFilterPos()'
• new variable "$dosage_alt" in 'seqGetData()' and 'seqApply()'
• import VCF files with no GT in 'seqVCF2GDS()'

UTILITIES

• 'seqDigest(f, "annotation/filter")' works on a factor variable
• improve the computational efficiency of 'seqMerge()' to avoid genotype recompression by padding the 2-bit genotype array in bytes
• significantly improve 'seqBlockApply()' (its speed is close to 'seqApply()')
• reduce the overhead in 'seqSetFilter(, variant.sel=...)'

SeqArray 1.18.2

BUG FIXES

• fix an issue: 'seqSetFilterChrom()' extends a genomic range upstream and downstream 1bp
• use '.onLoad()' instead of '.onAttach()' to fix https://support.bioconductor.org/p/104405/#104443

SeqArray 1.18.0

NEW FEATURES

• progress information: showing overall running time when completed
• new variable names "$ref" and "$alt" can be used in 'seqGetData()' and 'seqBlockApply()'
• new argument '.progress' in 'seqDigest()'
• new argument 'ref.allele' in 'seqAlleleCount()'
• new variable name "$chrom_pos_allele" can be used in 'seqGetData()' and 'seqBlockApply()'

UTILITIES

• move VariantAnnotation to the suggest field from the import field
• remove an unused argument '.list_dup' in 'seqBlockApply()'
• slightly improve the computational efficiency of ‘seqAlleleFreq()' and 'seqAlleleCount()' when ’ref.allele=0'
• ‘seqGetData(f, "$chrom_pos")' outputs characters with the format ’chromosome:position' instead of 'chromosome_position'

BUG FIXES

• fix the unexpected behaviors in 'seqSetFilter(, action="push")' and 'seqSetFilter(, action="push+intersect")'
• fix a bug in 'seqGetData(f, "$dosage")' when the number of unique alleles at a site greater than 3 (https://github.com/zhengxwen/SeqArray/issues/21)
• fix a bug in 'seqSNP2GDS()' for inverted genotypes during importing data from SNP GDS files (https://github.com/zhengxwen/SeqArray/issues/22)
• fix an issue of no phase data in 'seqExport()'

SeqArray 1.16.0

• a new argument 'intersect' in 'seqSetFilter()' and 'seqSetFilterChrom()'
• a new function 'seqSetFilterCond()'
• 'seqVCF2GDS()' allows arbitrary numbers of different alleles if REF and ALT in VCF are missing
• optimize internal indexing for FORMAT annotations to avoid reloading the indexing from the GDS file
• a new CITATION file
• 'LZMA_RA' is the default compression method in 'seqBED2GDS()' and 'seqSNP2GDS()'
• 'seqVCF_Header()' correctly calculates ploidy with missing genotypes

SeqArray 1.14.1

• The default compression setting in 'seqVCF2GDS()' and 'seqMerge()' is changed from "ZIP_RA" to "LZMA_RA"
• 'seqVCF2GDS()': variable-length encoding method is used to store integers in the FORMAT field of VCF files to reduce the file size and compression time

SeqArray 1.12.9

• the version number was bumped for the Bioconductor release version 3.3
• 'seqVCF_SampID()', 'seqVCF_Header()' and 'seqVCF2GDS()' allow a connection object instead of a file name
• "$num_allele" is allowed in 'seqGetData()' and 'seqApply()' (the numbers of distinct alleles)
• a new option '.progress' in 'seqAlleleFreq()', 'seqMissing()' and 'seqAlleleCount()'
• 'as.is' can be a 'gdsn.class' object in 'seqApply()'
• v1.12.7: a new argument 'parallel' in 'seqApply()', BiocParallel integration in 'seqParallel()' and a new function 'seqBlockApply()'
• v1.12.8: a new function 'seqGetParallel()'

SeqArray 1.12.0

• utilizes the official C API 'R_GetConnection()' to accelerate text import and export, requiring R (>=v3.3.0); alternative version (backward compatible with R_v2.15.0) is also available on github (https://github.com/zhengxwen/SeqArray/releases/tag/v1.11.18)
• ~4x speedup in the sequential version of 'seqVCF2GDS()', and 'seqVCF2GDS()' can run in parallel
• variables in "annotation/format/" should be two-dimensional as what mentioned in the vignette.
• rewrite 'seqSummary()'
• a new vignette file with Rmarkdown format (replacing SeqArray-JSM2013.pdf)
• bug fix in 'seqBED2GDS()' if the total number of genotypes > 2^31 (integer overflow)
• bug fixes in 'seqMerge()' if chromosome and positions are not unique
• 'seqStorage.Option()' is renamed to 'seqStorageOption()'
• new function 'seqDigest()'
• 'seqVCF.Header()' is renamed to 'seqVCF_Header()', 'seqVCF.SampID()' is renamed to 'seqVCF_SampID()'
• seqSetFilter(): 'samp.sel' is deprecated since v1.11.12, please use 'sample.sel' instead
• accelerate reading genotypes with SSE2(+13%) and AVX2(+23%)
• new function 'seqSystem()'
• allow "$dosage" in 'seqGetData()' and 'seqApply()' for the dosages of reference allele
• accelerate 'seqSetFilterChrom()' and allow a selection with multiple regions
• new methods '\S4method{seqSetFilter}{SeqVarGDSClass, GRanges}()' and '\S4method{seqSetFilter}{SeqVarGDSClass, GRangesList}()'
• 'as.is' in ‘seqApply()' allows a ’connection' object (created by file, gzfile, etc)
• 'seqSummary(f, "genotype")$seldim' returns a vector with 3 integers (ploidy, # of selected samples, # of selected variants) instead of 2 integers

SeqArray 1.10.6

• fix a memory issue in ‘seqAlleleFreq()' when ’ref.allele' is a vector
• ‘seqSetFilter()' allows numeric vectors in ’samp.sel' and 'variant.sel'
• 'seqSummary()' returns ploidy and reference
• 'seqStorage.Option()' controls the compression level of FORMAT/DATA
• ‘seqVCF2GDS()' allows extract part of VCF files via ’start' and 'count'
• 'seqMerge()' combines multiple GDS files with the same samples
• export methods for compatibility with VariantAnnotation
• a new argument '.useraw' in 'seqGetFilter()'
• a new argument 'allow.duplicate' in 'seqOpen()'
• fix a bug in 'seqParallel()' (https://github.com/zhengxwen/SeqArray/issues/11) and optimize its performance
• 'gdsfile' could be NULL in 'seqParallel()'

SeqArray 1.10.0

• a new function 'seqGDS2SNP()'
• supported by the SNPRelate package
• support 'seqApply(..., margin="by.sample")'
• new functions 'seqOptimize()', 'seqMissing()', 'seqAlleleFreq()', 'seqNumAllele()' and 'seqSetFilterChrom()'
• "intersection" and "push+intersection" in 'seqSetFilter()'
• parallel implementation in 'seqNumAllele()', 'seqMissing()' and 'seqAlleleFreq()'
• a new function 'seqExport()'
• new argument ".useraw" in 'seqApply()'
• fix a bug for duplicated "variant.id", https://github.com/zhengxwen/SeqArray/issues/7
• fix an issue of 'seqVCF2GDS()' when there are duplicated format or info ID
• improve access speed (+50%, benchmark on calling seqApply(..., FUN=function(x) {}))
• new functions 'seqSNP2GDS()', 'seqBED2GDS()', 'seqAlleleCount()' and 'seqResetFilter()'
• 'seqCompress.Option()' is renamed to 'seqStorage.Option()'
• "ZIP_RA" is the default value in 'seqStorageOption()' and other functions instead of "ZIP_RA.max"
• 'seqSetFilter()' becomes a S4 method

SeqArray 1.8.0

• bug fix in getting genotypes if position > 2^31
• add an option 'ignore.chr.prefix' to the function 'seqVCF2GDS()'
• 'seqVCF2GDS()' ignores the INFO or FORMAT variables if they are not defined ahead
• a new action 'push+set' in the function 'seqSetFilter()'
• bug fix if 'requireNamespace("SeqArray")' is called from other packages

SeqArray 1.6.0

• fix a bug in 'seqVCF2GDS()' when the values in the FILTER column are all missing
• enhance 'seqVCF.Header()'
• support the LinkingTo mechanism
• fix the error in haploid genotypes (Y chromosome)

SeqArray 1.4.0

• update according to the new version of VariantAnnotation
• update test codes to avoid the conflict
• bumped version as all packages that depend on Rcpp must be rebuilt
• modify to new biocViews to DESCRIPTION file

SeqArray 1.2.0

• add a new argument "action" to the function 'seqSetFilter()'
• add a new function 'seqInfoNewVar' which allows adding new variables to the INFO fields
• minor bug fix in asVCF
• update man page "SeqVarGDSClass-class.Rd" with new methods
• in DESCRIPTION, BiocGenerics listed in "Suggests" instead of "Imports" as suggested by R CMD check
• bug fix in seqDelete
• revise the function 'seqTranspose' according to the update of gdsfmt (v1.0.0)
• revise the argument 'var.index' in the function 'seqApply()'
• basic supports of 'GRanges' and 'DNAStringSetList'
• added methods 'qual', 'filt', 'asVCF'
• 'granges' method uses length of reference allele to set width
• minor bug fix to avoid 'seqGetData()' crashing when no value returned from a variable-length variable
• update documents

SeqArray 1.0.0

• the version number was bumped for the Bioconductor release version

SeqArray 0.99.0

• initial Bioconductor package submission