October 7, 2016
bam-load: 10x Genomics CB and UB tags are preserved
bam-load: Orphaned secondary alignments will be converted to primary alignments
bam-load: READ_GROUP is populated from 'BC' if 'RG' has no value
bam-load: fixed support for '-V' and '--version' command-line options
blast: Updated BLAST engine to 2.5.0+ version
build, ngs-tools: Now ngs-tools look for its dependencies using their normal build paths and does not reconfigure them
build, ngs-tools: Now ngs-tools use CMAKE_INSTALL_PREFIX for installation path
build, sra-tools: Now makefile copies default.kfg file when it is updated
build, sra-tools: separate decryption package was removed - decryption tools are included as part of sratoolkit.
kfg, kns: Use environment variables to determine the proxy for a http protocol
kfg, vdb-config: vdb-config-GUI has now option to pick up proxy-setting from environment
kns: All tools and libraries now support https
kns: replaced all direct uses of sleep() within code to enforce standardization upon mS.
kproc, ncbi-vdb: Fixed KCondition to generate timeout error on Windows when timeout exhausted
latf-load: now handles column values up to 64MB long
ngs: Fixed all crashes when using null as string in ngs-java APIs
ngs: NGS_ReferenceGetChunk() will now return chunks potentially exceeding 5000 bases
ngs: fixed potential concurrency issues at exit, when called from Java
ngs: ngs-java and ngs-python auto-download (of native libraries) now works through HTTPS
ngs: read fragments of length 0 are now ignored
ngs, ngs-tools, ref-variation: added class ngs-vdb::VdbAlignment, featuring method IsFirst()
ngs-engine: improved diagnostic messages
ngs-tools: Fixed Makefiles to keep supporting "./configure; make" build of sra-search, alongside CMake-based build.
prefetch: Fixed prefetch not to print misleading 'unknown integer storage type' error messages
sam-dump: CB and UB tags are now created if loaded via bam-load from 10xSingleCell
sra-tools: presence of ./ncbi (even if empty) subdirectory next to the executable files is no longer required, unless configuration files are needed.
test: updated tests to not fail outside of NCBI
test-sra: test-sra prints network information
test-sra: test-sra prints version of ncbi-vdb or ngs-sdk dynamic library
vdb: improved parameter checking on VDatabaseOpenTableRead()
vdb: new function: "VDBManagerDeleteCacheOlderThan()"
vdb: problem with buffer-overrun when compressing random data fixed
vdb: remote/aux nodes have been removed from configuration
vdb-dump: does not ignore table-argument on plain table any more, has to be SEQUENCE on plain tables if used
June 12, 2016
align, bam-load: Insert-only alignments no longer cause incorrect binning
bam-load: fixed case where WGS accessions where not being read correctly
bam-load: will NOT perform spot assembly using hard clipped secondary alignments, even when 'make-spots-with-secondary' is enabled; WILL perform spot assembly using hard-clipped secondary alignments when 'defer-secondary' is enabled
blast, kfg, ncbi-vdb, sra-tools, vfs: restored possibility to disable local caching
build, sra-tools: Running make and make install in sra-tools repository prepares all configuration files required to access NCBI repository
doc, ncbi-vdb: created a Wiki page illustrating how to use the API to set up logging
fastdump, sra-tools: new tool to perform fast dumps of a whole run in either fastq or a custom format for one of our customers.
kar: Alter the default ordering of components of an SRA archive for better network performance
kdb, kfs, kns: Added blob validation for data being fetched remotely
kfg: When loading configuration files on Windows USERPROFILE environment variable is used before HOME
kfg: modified auxiliary remote access configuration to use load-balanced servers
kns: Fixed a bug when KHttpRequestPOST generated an incorrect Content-Length after retry
ngs, search, sra-search: sra-search was modified to support multiple threads.
ngs-engine, ngs-tools, sra-tools, vfs: The "auxiliary" nodes in configuration are now ignored
pileup-stats: updated commandline parser to eat unprocessed parameters
sam-dump: updated to append asterisks to quality field of SAM output when corrupt original BAM has secondary alignment is shorter than the primary.
search: now supports multi-threaded search
sra-search: now supports sorted output
sra-tools: added possibility to build rpm package in sra-toolkit
sra-tools: fixed exit codes for a number of applications in response to command line options
vdb-dump: added tests to verify vdb-dump operation on nested databases
vdb-validate: A new checks were added for SEQUENCE table
vdb: fixed a bug in VCursorFindNextRowIdDirect where it returned a wrong rowId
vdb: fixed a bug in the code used to iterate over blobs where rowmap expansion cache would reset iteration to initial row instead of respecting sequence
vfs: environment variable VDB_PWFILE is no longer used
May 25, 2016
bam-load: Corrects an optimization used to compare read lengths when lengths are greater than 255 bam-load: alignments which are below the minimum match count but with at least 1/2 of the aligned bases matching are accepted bam-load: improved performance of SAM parsing code bam-load: non-fatal result codes no longer cause the reader thread to quit early bam-load: will NOT do spot assembly using hard clipped secondary alignments even when make-spots-with-secondary is enabled; WILL do spot assembly using hard-clipped secondary alignments when defer-secondary is enabled build: MSVS 2013 toolset (12.0) is now supported across all repositories vdb: Fixed a bound on memory cache that would never flush under certain access modes
April 20, 2016
align-cache: a tool producing vdbcache that reduces effects of random access and improves speed.
bam-load: Fixed bug in validation code for unsorted bam files
bam-load: If two (or more) local reference names refer to the same global reference, bam-load will record the first one used and report the change.
bam-load: Secondary alignment will be used for spot assembly if the loader is explicitly told to use them
bam-load: The code that had invalid item in index problem was removed in the process of addressing a performance issue.
bam-load: change reporting of fatal warnings into fatal errors
bam-load: changed from an error to a warning if using secondary alignments to create spots
bam-load: low-match secondary alignments are now discarded; low-match primary alignments are logged, and if too many, it will abort the load.
bam-load: records the contents of the BX tag
bam-load: rules for spot assembly were reverted to assembling spots only on primary alignments
blast, build: Improved blast tools: all required libraries are linked statically.
build: Allow to build sra-tools on systems without static c++ library
check-corrupt, vdb-validate: Added a new set of checks that can be triggered by using one of two "--sdc:" cmd options
copycat: now runs on Centos 7; no longer uses system-installed magic file
dbgap-mount: Added support for standard options "-L" and "-o", which allow users to determine the logging level and log output file
dbgap-mount: New optional parameter was introduced '-u' which allows user to unmount FUSE/DOKAN volume. Implemented on linux and windows
fastq-load: The (old) fastq-loader will properly report multiple reads on the 454 platform (and still fail).
kar: added '--md5' option to create md5sum compatible auxiliary file
kdb: Fixed VTableDropColumn, so that it can drop static columns
kfs, kns, ngs, sra-tools: Fixed thread safety issues for both cache and http files
kget: has a new option --full to match wget in speed. added examples.sh
kproc: Fixed KQueue to wake waiters when sealed, fixed KSemaphore to wake waiters when canceled
latf-load: now allows undescores inside spot group names
latf-load: now loads data produced by fastq-dump
latf-load: updated to support Illumina tag line format with identifier at the front
pileup-stats: added -V (--version) option: prints out the software
pileup-stats: pileup-stats: added version support (options -V or --version)
prefetch: Added --eliminate-quals option which speeds up HTTP download by ignoring QUALITY column`s data
prefetch: Fixed failure when running prefetch.exe "-a<bin|key>" when there is a space after "-a"
prefetch: messages about maximum size of download are made more user-friendly.
prefetch: now will download even when caching is disabled
ref-variation: --input-file option allows to specify input accessions and paths in the file
ref-variation: added "count-strand" option: it controls relative orientation of 3' and 5' fragments.
ref-variation: added -c option to flush output immediately; reporting zero matches
ref-variation: added a way to specify a number of repeats of the query
ref-variation: improved threading management
ref-variation: removed irrelevant warnings reported in some cases in debug version only
sam-dump: Segfault no longer occurs when confronted with large amounts of header lines
sam-dump: added option to produce MD tags
sam-dump: filters out duplicates in the rows that it generates
sam-dump: produces BX-tags if preserved in SRA file by bam-load
sra-sort: correctly generates spot-id column even in the absence of primary alignments
sra-stat: no longer fails when CS_NATIVE column is not present.
sra-tools, vdb-config: Removed dependency of mac binaries on unnecessary libraries, e.g. libxml2.
sra-tools: ncbi#27 : contains short and long examples of how to configure sra-tools build
var-expand: a tool for batch variation expansion
vdb-config: now handles standard options such as --option-file
vdb-validate: Added code to continue with the next row when column has discontiguous blobs