NEWS

CHANGES IN VERSION 1.46.1
-------------------------

UTILITIES

    o `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.


CHANGES IN VERSION 1.46.0
-------------------------

UTILITIES

    o `seqGetData()` return NULL, if 'var.name=character()'


CHANGES IN VERSION 1.44.3
-------------------------

UTILITIES

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


CHANGES IN VERSION 1.44.2
-------------------------

BUG FIXES

    o fix `seqAddValue(, val=vector("list", NUM_VARIANT))`

    o fix the ploidy returned from `seqVCF_Header()`, when there are genotypes
      of males and females on Chromosome X


CHANGES IN VERSION 1.44.1
-------------------------

UTILITIES

    o new option 'numvariant' in `seqEmptyFile()`

BUG FIXES

    o `seqMerge()` should internally use "chr_position_ref_alt" to distinguish
      the variants in different files

    o `seqAddValue(, varnm="annotation/filter")` should work with a factor
      variable

    o `seqAddValue(, varnm="variant.id")` can reset the variant IDs with a
      different number of the variants


CHANGES IN VERSION 1.44.0
-------------------------

UTILITIES

    o tweak the display of progress information in `seqVCF2GDS()`

    o `seqVCF_Header(, getnum=TRUE, verbose=TRUE)` to show the progress
      information for scanning the VCF file

    o 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)

    o new 'verbose.clean' in `seqExport()` to control how much information to
      be displayed


CHANGES IN VERSION 1.42.4
-------------------------

BUG FIXES

    o `seqGetData(, "$dosage_alt")` and `seqGetData(, "$dosage_sp")` work
      correctly when the ploidy is >2 and there are missing alleles

    o fix a bug that `seqParallel()` does not call a user-defined '.combine'
      when 'parallel=1'


CHANGES IN VERSION 1.42.1
-------------------------

UTILITIES

    o update the help files of `seqBlockApply()` and `seqUnitApply()`

    o detect the output filename extension in `seqGDS2VCF()` without
      considering the case of the characters, supporting .gz, .bgz, .bz and .xz
      as a filename extension

    o fix the compiler warning: -Wformat-security

    o new option 'include.pheno=TRUE' in `seqBED2GDS()`


CHANGES IN VERSION 1.42.0
-------------------------

UTILITIES

    o new option 'write.rsid' in `seqGDS2BED()`


CHANGES IN VERSION 1.40.1
-------------------------

BUG FIXES

    o `seqAddValue(gdsfile, varnm="position")` works correctly


CHANGES IN VERSION 1.40.0
-------------------------

    o fix the compiler warning: sprintf is deprecated


CHANGES IN VERSION 1.38.0
-------------------------

UTILITIES

    o new option 'ext_nbyte' in `seqGet2bGeno()`

    o `seqAlleleCount()` and `seqGetAF_AC_Missing()` return NA instead of zero
      when all genotypes are missing at a site

    o `seqGDS2VCF()` does not output the FORMAT column if there is no selected
      sample (e.g., site-only VCF files)

    o `seqGetData(, "$chrom_pos2")` is similar to `seqGetData(, "$chrom_pos")`
      except the duplicates with the suffix ("_1", "_2" or >2)

NEW FEATURES

    o `seqGDS2BED()` can convert to PLINK BED files with the best-guess
      genotypes when there are only numeric dosages in the GDS file

    o `seqEmptyFile()` outputs an empty GDS file


CHANGES IN VERSION 1.36.2
-------------------------

BUG FIXES

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


CHANGES IN VERSION 1.36.1
-------------------------

BUG FIXES

    o `seqExport()` failed when there is no variant

    o `seqSetFilter(, ret.idx=TRUE)`, see
      https://github.com/zhengxwen/SeqArray/issues/80


CHANGES IN VERSION 1.36.0
-------------------------

NEW FEATURES

    o new functions `seqUnitCreate()`, `seqUnitSubset()` and `seqUnitMerge()`

    o new functions `seqFilterPush()` and `seqFilterPop()`

    o new functions `seqGet2bGeno()` and `seqGetAF_AC_Missing()`

    o new function `seqGetData(, "$dosage_sp")` for a sparse matrix of dosages

    o the first argument 'gdsfile' can be a file name in `seqAlleleFreq()`,
      `seqAlleleCount()`, `seqMissing()`

    o new function `seqMulticoreSetup()` for setting a multicore cluster
      according to a numeric value assigned to the argument 'parallel'

UTILITIES

    o 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()`

    o remove the deprecated '.progress' in `seqMissing()`, `seqAlleleCount()`
      and `seqAlleleFreq()`

    o add `summary.SeqUnitListClass()`

    o no genotype and phase data nodes from `seqSNP2GDS()` if SNP dosage GDS
      is the input

BUG FIXES

    o `seqUnitApply()` works correctly with selected samples if 'parallel' is
      a non-fork cluster

    o `seqVCF2GDS()` and `seqVCF_Header()` work correctly if the VCF header has
       white space

    o `seqGDS2BED()` with selected samples for sex and phenotype information

    o `seqGDS2VCF()` failed if there is no 'genotype/data' in the GDS file


CHANGES IN VERSION 1.32.0
-------------------------

NEW FEATURES

    o new option 'ret.idx' in `seqSetFilter()` for unsorted sample and variant
      indices

    o new option 'ret.idx' in `seqSetFilterAnnotID()` for unsorted variant
      index

    o rewrite the function `seqSetFilterPos()`: new options 'ref' and 'alt',
      'multi.pos=TRUE' by default

    o new option 'packed.idx' in `seqAddValue()` for packing an indexing
      variable

    o new option 'warn' in `seqSetFilter()` to enable or disable the warning

    o new functions `seqNewVarData()` and `seqListVarData()` for
      variable-length data

UTILITIES

    o allow no variant in `seqApply()` and `seqBlockApply()`

    o the list object returned from `seqGetData()` always have names if there
      are more than one input variable names

BUG FIXES

    o `seqGDS2VCF()` should output "." instead of NA in the FILTER column

    o `seqGetData()` should support factor when '.padNA=TRUE' or '.tolist=TRUE'
    
    o fix `seqGDS2VCF()` with factor variables

    o `seqSummary(gds, "$filter")` should return a data frame with zero row if
      'annotation/filter' is not a factor


CHANGES IN VERSION 1.30.0
-------------------------

UTILITIES

    o show a warning when an unsorted index is used in `seqSetFilter()`

    o show a message if `seqVCF_Header()` fails

    o a new option 'chr_prefix' in `seqGDS2VCF()`

BUG FIXES

    o `seqVCF_Header()` fixes 'contig' in the header of VCF if there are
      different fields


CHANGES IN VERSION 1.28.1
-------------------------

BUG FIXES

    o `seqRecompress(, verbose=FALSE)` works correctly

    o `seqSetFilter(, action="push+set")` should not reset the filter
      before setting a new filter


CHANGES IN VERSION 1.28.0
-------------------------

NEW FEATURES

    o new functions `seqUnitSlidingWindows()`, `seqUnitApply()`,
      `seqUnitFilterCond()`

    o new variable "$variant_index", "$sample_index" in `SeqGetData()`,
      `seqBlockApply()` and `seqUnitApply()` to get the indices of selected
      variants

    o new arguments '.padNA' and '.envir' in `seqGetData()`

    o new functions `seqSetFilterAnnotID()` and `seqGDS2BED()`

    o multicore function in `seqBED2GDS(, parallel=)`

    o new package-wide option `options(seqarray.nofork=TRUE)` to disable forking

    o new option 'minor' in `seqAlleleFreq()` and `seqAlleleCount()`

    o new option 'verbose' in `seqMissing()`, `seqAlleleFreq()` and
      `seqAlleleCount()`; '.progress' is deprecated, but still can be used
      for compatiblity

    o `seqAlleleFreq()`, `seqAlleleCount()`, `seqMissing()`,
      `seqSetFilterCond()` work on 'annotation/format/DS', if 'genotype/data'
      is not available

UTILITIES

    o `seqAddValue()` adds vectors, matrices and data frame to "annotation/info"

    o `seqBED2GDS()` allows a single file name without the extended file names
      (.bed, .fam, .bim)

    o allele flip in `seqBED2GDS()` to allow major allele to be reference

    o rewrite `seqGetData()` for faster loading

    o significantly improve `seqBlockApply()` on 'annotation/info/VARIABLE'
      (https://github.com/zhengxwen/SeqArray/issues/59)

    o add a S3 method `print.SeqVCFHeaderClass()` for `seqVCF_Header()`

    o new option '.tolist' in `seqGetData()`, `seqBlockApply()` and
      `seqUnitApply()`

    o String "." in a VCF file are converted to a blank string (missing value)
      in `seqVCF2GDS()`

    o add a class name 'SeqVarDataList' to the returned 'list(length, data)'
      from `seqGetData()`

    o new option `seqMissing(, per.variant=NA)`

    o add `comment.char=""` to `seqBED2GDS()`


CHANGES IN VERSION 1.26.2
-------------------------

NEW FEATURES

    o multiple variable names are allowed in `seqGetData(, var.name=)`

BUG FIXES

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


CHANGES IN VERSION 1.26.1
-------------------------

BUG FIXES

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


CHANGES IN VERSION 1.26.0
-------------------------

NEW FEATURES

    o new function `seqAddValue()`

UTILITIES

    o RLE chromosome coding in `seqBED2GDS()`

    o change the file name "vignettes/R_Integration.Rmd" to
      "vignettes/SeqArray.Rmd", so `vignette("SeqArray")` can work directly

    o correct Estimated remaining Time to Complete (ETC) for load balancing in
      `seqParallel()`

BUG FIXES

    o `seqBED2GDS(, verbose=FALSE)` should have no display

CHANGES

    o use a svg file instead of png in vignettes


CHANGES IN VERSION 1.24.2
-------------------------

NEW FEATURES

    o add the compiler information in `seqSystem()`

    o new arguments '.balancing', '.bl_size' and '.bl_progress' in
      `seqParallel()` for load balancing

UTILITIES

    o improve unix forking processes for load balancing in `seqParallel()`

BUG FIXES

    o fix `seqSummary()` when no phase data


CHANGES IN VERSION 1.24.0
-------------------------

NEW FEATURES

    o a new function `seqResetVariantID()`

    o a new option in `seqRecompress(, compress="none")` to uncompress all data

    o `seqGetData()` allows a GDS file name in the first argument


CHANGES IN VERSION 1.22.6
-------------------------

BUG FIXES

    o `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)

    o the bgzf VCF file is truncated in `seqGDS2VCF()` since the file is not
      closed appropriately

    o invalid chromosomes and position in the output of `seqMerge()` when
      merging different samples but same variants


CHANGES IN VERSION 1.22.3
-------------------------

NEW FEATURES

    o a new option 'scenario' in `seqVCF2GDS()` and `seqBCF2GDS()`

UTILITIES

    o more information in `seqDelete()`

BUG FIXES

    o export a haploid VCF file using `seqGDS2VCF()`

    o export VCF without any FORMAT data in `seqGDS2VCF()`

    o export GDS without genotypes in `seqExport()`

    o fix parallel file writing in seqVCF2GDS(), when no genotype


CHANGES IN VERSION 1.22.0
-------------------------

NEW FEATURES

    o `seqSNP2GDS()` imports dosage GDS files

    o `seqVCF_Header()` allows a BCF file as an input

    o a new function `seqRecompress()`

    o a new function `seqCheck()` for checking the data integrity of a SeqArray
      GDS file

    o `seqGDS2SNP()` exports dosage GDS files

UTILITIES

    o avoid duplicated meta-information lines in `seqVCF2GDS()` and
      `seqVCF_Header()`

    o require >= R_v3.5.0, since reading from connections in text mode is
      buffered

    o `seqDigest()` requires the digest package

    o optimization in reading genotypes from a subset of samples (according to
      gdsfmt_1.17.5)

BUG FIXES

    o `seqVCF2GDS()` and `seqVCF_Header()` are able to import site-only VCF
      files (i.e., VCF with no sample)

    o fix `seqVCF2GDS()` and `seqBCF2GDS()` since reading from connections in
      text mode is buffered in R >= v3.5.0


CHANGES IN VERSION 1.20.1
-------------------------

BUG FIXES

    o `seqExport()` fails to export haploid data (e.g., Y chromosome)

    o `seqVCF2GDS()` fails to convert INFO variables when Number="R"


CHANGES IN VERSION 1.20.0
-------------------------

NEW FEATURES

    o `seqGDS2VCF()` outputs a bgzip vcf file for tabix indexing

    o two more options "Ultra" and "UltraMax" in `seqStorageOption()`

    o '@chrom_rle_val' and '@chrom_rle_len' are added to a GDS file for
      faster chromosome indexing

    o new function `seqBCF2GDS()` (requiring the software bcftools)

    o new function `seqSetFilterPos()`

    o new variable "$dosage_alt" in `seqGetData()` and `seqApply()`

    o import VCF files with no GT in `seqVCF2GDS()`

UTILITIES

    o `seqDigest(f, "annotation/filter")` works on a factor variable

    o improve the computational efficiency of `seqMerge()` to avoid genotype
      recompression by padding the 2-bit genotype array in bytes

    o significantly improve `seqBlockApply()` (its speed is close to
      `seqApply()`)

    o reduce the overhead in `seqSetFilter(, variant.sel=...)`


CHANGES IN VERSION 1.18.2
-------------------------

BUG FIXES

    o fix an issue: `seqSetFilterChrom()` extends a genomic range
      upstream and downstream 1bp

    o use `.onLoad()` instead of `.onAttach()` to fix
      https://support.bioconductor.org/p/104405/#104443


CHANGES IN VERSION 1.18.0
-------------------------

NEW FEATURES

    o progress information: showing overall running time when completed

    o new variable names "$ref" and "$alt" can be used in `seqGetData()` and
      `seqBlockApply()`

    o new argument '.progress' in `seqDigest()`

    o new argument 'ref.allele' in `seqAlleleCount()`

    o new variable name "$chrom_pos_allele" can be used in `seqGetData()` and
      `seqBlockApply()`

UTILITIES

    o move VariantAnnotation to the suggest field from the import field

    o remove an unused argument '.list_dup' in `seqBlockApply()`

    o slightly improve the computational efficiency of `seqAlleleFreq()` and
      `seqAlleleCount()` when 'ref.allele=0'

    o `seqGetData(f, "$chrom_pos")` outputs characters with the format
      'chromosome:position' instead of 'chromosome_position'

BUG FIXES

    o fix the unexpected behaviors in `seqSetFilter(, action="push")` and
      `seqSetFilter(, action="push+intersect")`

    o 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)

    o fix a bug in `seqSNP2GDS()` for inverted genotypes during importing data
      from SNP GDS files (https://github.com/zhengxwen/SeqArray/issues/22)

    o fix an issue of no phase data in `seqExport()`


CHANGES IN VERSION 1.16.0
-------------------------

    o a new argument 'intersect' in `seqSetFilter()` and `seqSetFilterChrom()`

    o a new function `seqSetFilterCond()`

    o `seqVCF2GDS()` allows arbitrary numbers of different alleles if REF and
      ALT in VCF are missing

    o optimize internal indexing for FORMAT annotations to avoid reloading
      the indexing from the GDS file

    o a new CITATION file

    o 'LZMA_RA' is the default compression method in `seqBED2GDS()` and
      `seqSNP2GDS()`

    o `seqVCF_Header()` correctly calculates ploidy with missing genotypes


CHANGES IN VERSION 1.14.1
-------------------------

    o The default compression setting in `seqVCF2GDS()` and `seqMerge()` is
      changed from "ZIP_RA" to "LZMA_RA"

    o `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


CHANGES IN VERSION 1.12.9
-------------------------

    o the version number was bumped for the Bioconductor release version 3.3

    o `seqVCF_SampID()`, `seqVCF_Header()` and `seqVCF2GDS()` allow a
      connection object instead of a file name

    o "$num_allele" is allowed in `seqGetData()` and `seqApply()` (the numbers
      of distinct alleles)

    o a new option '.progress' in `seqAlleleFreq()`, `seqMissing()` and
      `seqAlleleCount()`

    o 'as.is' can be a `gdsn.class` object in `seqApply()`

    o v1.12.7: a new argument 'parallel' in `seqApply()`, BiocParallel
      integration in `seqParallel()` and a new function `seqBlockApply()`

    o v1.12.8: a new function `seqGetParallel()`


CHANGES IN VERSION 1.12.0
-------------------------

    o 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)

    o ~4x speedup in the sequential version of `seqVCF2GDS()`, and
      `seqVCF2GDS()` can run in parallel

    o variables in "annotation/format/" should be two-dimensional as what
      mentioned in the vignette.

    o rewrite `seqSummary()`

    o a new vignette file with Rmarkdown format (replacing SeqArray-JSM2013.pdf)

    o bug fix in `seqBED2GDS()` if the total number of genotypes > 2^31
      (integer overflow)

    o bug fixes in `seqMerge()` if chromosome and positions are not unique

    o `seqStorage.Option()` is renamed to `seqStorageOption()`

    o new function `seqDigest()`

    o `seqVCF.Header()` is renamed to `seqVCF_Header()`,
      `seqVCF.SampID()` is renamed to `seqVCF_SampID()`

    o seqSetFilter(): 'samp.sel' is deprecated since v1.11.12, please use
      'sample.sel' instead

    o accelerate reading genotypes with SSE2(+13%) and AVX2(+23%)

    o new function `seqSystem()`

    o allow "$dosage" in `seqGetData()` and `seqApply()` for the dosages of
      reference allele

    o accelerate `seqSetFilterChrom()` and allow a selection with
      multiple regions

    o new methods `\S4method{seqSetFilter}{SeqVarGDSClass, GRanges}()` and
      `\S4method{seqSetFilter}{SeqVarGDSClass, GRangesList}()`

    o 'as.is' in `seqApply()` allows a 'connection' object (created by file,
      gzfile, etc)

    o `seqSummary(f, "genotype")$seldim` returns a vector with 3 integers
      (ploidy, # of selected samples, # of selected variants) instead of
      2 integers


CHANGES IN VERSION 1.10.6
-------------------------

    o fix a memory issue in `seqAlleleFreq()` when 'ref.allele' is a vector

    o `seqSetFilter()` allows numeric vectors in 'samp.sel' and 'variant.sel'

    o `seqSummary()` returns ploidy and reference

    o `seqStorage.Option()` controls the compression level of FORMAT/DATA

    o `seqVCF2GDS()` allows extract part of VCF files via 'start' and 'count'

    o `seqMerge()` combines multiple GDS files with the same samples

    o export methods for compatibility with VariantAnnotation

    o a new argument '.useraw' in `seqGetFilter()`

    o a new argument 'allow.duplicate' in `seqOpen()`

    o fix a bug in `seqParallel()`
      (https://github.com/zhengxwen/SeqArray/issues/11) and optimize
      its performance

    o 'gdsfile' could be NULL in `seqParallel()`


CHANGES IN VERSION 1.10.0
-------------------------

    o a new function `seqGDS2SNP()`

    o supported by the SNPRelate package

    o support `seqApply(..., margin="by.sample")`

    o new functions `seqOptimize()`, `seqMissing()`, `seqAlleleFreq()`,
      `seqNumAllele()` and `seqSetFilterChrom()`

    o "intersection" and "push+intersection" in `seqSetFilter()`

    o parallel implementation in `seqNumAllele()`, `seqMissing()` and
      `seqAlleleFreq()`

    o a new function `seqExport()`

    o new argument ".useraw" in `seqApply()`

    o fix a bug for duplicated "variant.id",
      https://github.com/zhengxwen/SeqArray/issues/7

    o fix an issue of `seqVCF2GDS()` when there are duplicated format or
      info ID

    o improve access speed (+50%, benchmark on calling
      seqApply(..., FUN=function(x) {}))

    o new functions `seqSNP2GDS()`, `seqBED2GDS()`, `seqAlleleCount()` and
      `seqResetFilter()`

    o `seqCompress.Option()` is renamed to `seqStorage.Option()`

    o "ZIP_RA" is the default value in `seqStorageOption()` and other
      functions instead of "ZIP_RA.max"

    o `seqSetFilter()` becomes a S4 method


CHANGES IN VERSION 1.8.0
-------------------------

    o bug fix in getting genotypes if position > 2^31

    o add an option 'ignore.chr.prefix' to the function `seqVCF2GDS()`

    o `seqVCF2GDS()` ignores the INFO or FORMAT variables if they are not
      defined ahead

	o a new action 'push+set' in the function `seqSetFilter()`

    o bug fix if 'requireNamespace("SeqArray")' is called from other packages


CHANGES IN VERSION 1.6.0
-------------------------

    o fix a bug in `seqVCF2GDS()` when the values in the FILTER column are
      all missing

    o enhance `seqVCF.Header()`

    o support the LinkingTo mechanism

    o fix the error in haploid genotypes (Y chromosome)


CHANGES IN VERSION 1.4.0
-------------------------

    o update according to the new version of VariantAnnotation

    o update test codes to avoid the conflict

    o bumped version as all packages that depend on Rcpp must be rebuilt

    o modify to new biocViews to DESCRIPTION file


CHANGES IN VERSION 1.2.0
-------------------------

    o add a new argument "action" to the function `seqSetFilter()`

    o add a new function 'seqInfoNewVar' which allows adding new variables
      to the INFO fields

    o minor bug fix in asVCF

    o update man page "SeqVarGDSClass-class.Rd" with new methods

    o in DESCRIPTION, BiocGenerics listed in "Suggests" instead of "Imports"
      as suggested by R CMD check

    o bug fix in seqDelete

    o revise the function 'seqTranspose' according to the update of
      gdsfmt (v1.0.0)

    o revise the argument 'var.index' in the function `seqApply()`

    o basic supports of 'GRanges' and 'DNAStringSetList'

    o added methods 'qual', 'filt', 'asVCF'

    o 'granges' method uses length of reference allele to set width

    o minor bug fix to avoid `seqGetData()` crashing when no value returned
      from a variable-length variable

    o update documents


CHANGES IN VERSION 1.0.0
-------------------------

    o the version number was bumped for the Bioconductor release version


CHANGES IN VERSION 0.99.0
-------------------------

    o initial Bioconductor package submission