CHANGES IN VERSION 1.34.0
-------------------------
o fix the compiler warning: sprintf is deprecated
CHANGES IN VERSION 1.32.0
-------------------------
o update the web links
CHANGES IN VERSION 1.30.1
-------------------------
o return a object of S3 class "snpgdsGRMClass" in `snpgdsGRM()` instead of
a list when `with.id=TRUE`
o fix a portable issue with Fortran character strings via
USE_FC_LEN_T & FCONE
CHANGES IN VERSION 1.24.0
-------------------------
o definition of IBS in the `snpgdsIBS()` help file
CHANGES IN VERSION 1.22.0
-------------------------
o 'allow.fork=TRUE' is the default in `snpgdsOpen()` since v1.21.2
o add `print.snpgdsPCAClass`, `print.snpgdsEigMixClass`,
`print.snpgdsPCASNPLoadingClass`, `print, snpgdsEigMixSNPLoadingClass`,
`print.snpgdsIBDClass`, `print.snpgdsDissClass`, `print.snpgdsIBSClass`
o fix a Win32 compiler issue with gcc-4.9
o update `snpgdsBED2GDS()` with comment.char=""
CHANGES IN VERSION 1.20.1
-------------------------
o Bug fix: eigenvalues and TraceXTX are correctly calculated in
`snpgdsPCA(, algorithm="randomized")`
CHANGES IN VERSION 1.20.0
-------------------------
o a leading tilde in the file path is allowed in `snpgdsGDS2BED()`
o change the file name "vignettes/SNPRelateTutorial.Rmd" to
"vignettes/SNPRelate.Rmd", so `vignette("SNPRelate")` can work directly
CHANGES IN VERSION 1.18.1
-------------------------
o support long vector in `snpgdsIBDSelection()`
CHANGES IN VERSION 1.18.0
-------------------------
o `snpgdsBED2GDS()` allows a single file name without the extended file
names (.bed, .fam, .bim)
CHANGES IN VERSION 1.16.0
-------------------------
o a new option 'useMatrix' for the packed symmetric matrix using the Matrix
package in `snpgdsIBDMoM()`, `snpgdsIBDKING()`, `snpgdsIBS()`,
`snpgdsIndivBeta()` and `snpgdsGRM()` to reduce memory usages
o fix a bug of missing sample and SNP IDs in the output of `snpgdsIndInb()`
o new option 'start.pos' in `snpgdsLDpruning()`
o new methods in `snpgdsIndInb()`: gcta1, gcta2, gcta3; progress
information is shown during running the function
o `snpgdsCombineGeno()` supports dosages
CHANGES IN VERSION 1.14.0
-------------------------
o the default compression is "LZMA_RA" in `snpgdsBED2GDS()`,
`snpgdsVCF2GDS()` and `snpgdsVCF2GDS_R()` for annotations
o support Intel C++ compiler with SSE2/AVX2
o allow interrupting requests in the calculation
o new method options in `snpgdsPairScore()`: GVH.major, GVH.minor,
GVH.major.only, GVH.minor.only
o force to use integers for 'snp.position' in `snpgdsCreateGeno()`
o unit tests for merging GRMs in `snpgdsMergeGRM()`
o the function `snpgdsSNPListStrand()` is merged to
`snpgdsSNPListIntersect()`, and it is removed from the package
o update `snpgdsSNPListIntersect()` and `snpgdsCombineGeno()` (work
correctly)
o replace INF/-INF by NaN in the output of `snpgdsIBDKING()`
CHANGES IN VERSION 1.12.0-1.12.2
-------------------------
o new arguments 'with.sample.id' and 'with.snp.id' in `snpgdsSNPRateFreq()`
o fix an issue in the C code 'LENGTH or similar applied to NULL object'
o a new option 'method="Jacquard"' in `snpgdsPairIBD()`
o `snpgdsGRM()` can output the GRM matrix to a GDS file
o a new function `snpgdsMergeGRM()` to merge multiple GRMs
CHANGES IN VERSION 1.10.0-1.10.2
-------------------------
o new functions `snpgdsAdmixPlot()` and `snpgdsAdmixTable()`
o `snpgdsPCASNPLoading()` and `snpgdsPCASampLoading()` support the eigen
results of `snpgdsEIGMIX()` allowing projecting new samples to the
existing coordinate
o `snpgdsFst()` provides W&C84 mean Fst together with weighted Fst
o a new argument 'outgds' in `snpgdsPCACorr()` allows exporting
correlations to a gds file
o a friendly warning is given when openning a SeqArray file with
`snpgdsOpen()`
o a new option "Corr" in `snpgdsGRM()` for scaled GRM
o progress information: showing overall running time when completed
o An unexpected exception in a thread could result in deadlock: the current
implementation shows error information and exits the R session
CHANGES IN VERSION 1.8.0
-------------------------
o add a new function `snpgdsIndivBeta()`
CHANGES IN VERSION 1.6.0-1.6.6
-------------------------
o fix an issue in `snpgdsVCF2GDS()` if sample.id has white space
o bug fix in `snpgdsPCASampLoading()` when the input is SeqArray GDS file
o improve `snpgdsGetGeno()`
o the version number was bumped for the Bioconductor release version 3.3
o new implement of thread pool
o bitwise intrinsics (SSE2/AVX2) to accelerate `snpgdsIBSNum()`,
`snpgdsIBS()`, `snpgdsIBDMoM()`, `snpgdsIBDKing()` (+50% to +300%)
o v1.6.4: bug fix in v1.6.3 (allele counting error with SSE2 implement)
o v1.6.5: `snpgdsGRM()`, renames the option "Visscher" to "GCTA",
new option 'dosage' in `snpgdsPairScore()`,
new function `plot.snpgdsPCAClass()`
CHANGES IN VERSION 1.4.0
-------------------------
o update GENEVA website address
o add a new function `snpgdsPairScore()`
o improve `snpgdsGRM()`
o support the SeqArray data format (requiring SeqArray >= v1.9.2)
o modification according to gdsfmt_1.5.7
o bug fix in `snpgdsLDMat()` when genotype data are from SeqArray
o bug fix: https://github.com/zhengxwen/SNPRelate/issues/14
o `snpgdsLDMat()` supports multiple threads and covariance
o `snpgdsPCA()`: non-computed eigenvalues are NaN to avoid misuse of
'eigenval' when `eigen.method="DSPEVX"`
CHANGES IN VERSION 1.2.0
-------------------------
o fix a bug in `snpgdsVCF2GDS()` when 'method="biallelic.only"'
o add `snpgdsVCF2GDS_R()` for the R implementation
o fix a bug in `snpgdsBED2GDS()` if 'family=TRUE'
o `snpgdsGDS2BED()` allows the file name of GDS
o improve `snpgdsSlidingWindow()`
o add an option 'ignore.chr.prefix' to the function `snpgdsVCF2GDS()`
o a new function `snpgdsHWE()`
o add "Fst estimation" to the vignette
o bug fix if `requireNamespace("SNPRelate")` is called from other packages
o `snpgdsPCA()` uses 'DSPEVX' to compute eigenvalues and
eigenvectors instead of 'DSPEV' if top eigenvalues are required only
(significant improvement on computing speed)
o the original Rnw vignette is replaced by a R Markdown vignette
o a new function `snpgdsPED2GDS()`
CHANGES IN VERSION 1.0.0
-------------------------
o The version number was bumped for the Bioconductor release version
CHANGES IN VERSION 0.99.1-0.99.3
-------------------------
o an option to create an integer snp.id when converting from PLINK
o a new function `snpgdsFst()` to estimate Fst
o minor fixes
CHANGES IN VERSION 0.99.0
-------------------------
o be a Bioconductor package
o fully support long vectors (>= R v3.0)
o >5x speedup in the function `snpgdsVCF2GDS()`
o SNP GDS format allows character-type chromosome codes
o add a new argument 'ref.allele' in `snpgdsVCF2GDS()`
o add new functions `snpgdsOpen()` and `snpgdsClose()`
o add a new function `snpgdsTranspose()` to transpose the genotypic matrix
o add a new function `snpgdsAlleleSwitch()` to switch alleles if needed
o add a new function `snpgdsApartSelection()`
o add a new function `snpgdsGEN2GDS()` to import Oxford GEN data
o use NA instead of 3 as missing value in `snpgdsGetGeno()`
o add a new argument 'snpfirstdim' in the function `snpgdsGDS2BED()`
o add a new argument 'with.id' in the functions `snpgdsSNPRateFreq()`
and `snpgdsSampMissRate()`
o return a numeric vector instead of data.frame in `snpgdsLDpair()`
o add estimating nine Jacquard's coefficients in `snpgdsIBDMLE()`
o fix the memory issues reported by valgrind
CHANGES IN VERSION 0.9.15-0.9.19
-------------------------
o improve `snpgdsIBDKING()` to avoid misleading arguments and examples
o improve `snpgdsIBDSelection()`
o the names of column 'sample1', 'sample2' and 'kinshipcoeff' are changed
to 'ID1', 'ID2' and 'kinship' in the output of `snpgdsIBDSelection()`
o avoid compiler issues when the version of R < v3.0
o minor fixes in `snpgdsIBDSelection()`: allow non-character-type sample
IDs and missing IBD coefficients
CHANGES IN VERSION 0.9.14:
-------------------------
o `snpgdsGDS2BED()` gives a warning if the allele information is missing
o add `snpgdsIBDKING()` (KING robust relationship inference in genome-wide
association studies)
o bug fixes in `snpgdsIBS()` and `snpgdsIBSNum()`
o some warnings in `snpgdsSummary()` are replaced by hints
o bug fix in `snpgdsVCF2GDS()`: duplicate SNP id when importing multiple
VCF files
o improve hint information in `snpgdsGDS2BED()`
CHANGES IN VERSION 0.9.10-0.9.13
-------------------------
o disable parallel functions in the Windows version of SNPRelate, until
the internal multi-thread funcitons are replaced by the functions
in win-pthreads
o fix compiling issues in the Windows version of SNPRelate
o add 'GetRNGstate' and 'PutRNGstate' to the C function `gnrDistPerm()`
o remove the upper limit of 200 chromosomes in `snpgdsVCF2GDS()` according
to scaffolds/contigs from a de novo assembly not chromosomes
o improve `snpgdsSummary()`
o bug fixes in `snpgdsPairIBD()`
o bug fix in `snpgdsVCF2GDS()`: to use "\t" as the field separator
character
o bug fixes according to gdsfmt v0.9.11
o improve the vignette
CHANGES IN VERSION 0.9.8-0.9.9
-------------------------
o adds a new argument 'family' to `snpgdsBED2GDS()`
o improves `snpgdsCutTree()`
o adds a new function `snpgdsDrawTree()`
o adds an argument 'option' to the function `snpgdsBED2GDS()`
o outputs chromosome codes according to `snpgdsOption()` in
`snpgdsGDS2BED()`
o add an alternative method extracting variants from a VCF file in
`snpgdsVCF2GDS()`
o supports multiple files in `snpgdsVCF2GDS()`
o add `snpgdsOption()`
o add `snpgdsHCluster()`
o add `snpgdsCutTree()`
o add `snpgdsDiss()`
CHANGES IN VERSION 0.9.1-0.9.7
-------------------------
o bug fix: an important bug of missing genotypes in `snpgdsVCF2GDS()`
o bug fix: a bug in `snpgdsGDS2BED()` (output 23, 24, ..., instead of
X, Y, XY)
o bug fix: a bug in `snpgdsCombineGeno()` (due to duplicate SNPs)
o `snpgdsLDpair()` provides haplotype frequency estimates
o `snpgdsLDMat()` allows specifying sliding window size
o add RUnit tests
o add a vignette
o bug fixes
o add functions to `snpgdsIBDMLE()` and `snpgdsIBDMoM()`
o add a new function `snpgdsVCF2GDS()`
CHANGES IN VERSION 0.9.0 (03-Aug-2011)
-------------------------
o first public release of SNPRelate