Latest improvements for QIAGEN CLC Assembly Cell

Improvements

• Improved the alignment quality of the clc_mapper tool by removing aligned ends with an alignment score of zero. As a result, some alignments will be shorter and may be filtered away because they no longer pass the minimum length fraction criterion.

Improvements

• Kmer indexing in clc_assembler_long now uses a minimizer-based kmer index, resulting in improved performance for polished reads.
• When using the option “-a” in clc_mapping_table, mismatches are now indicated by dots in the alignment output instead of |.
• The following tools now support the “-h” option for listing help information: clc_assembler_long, clc_correct_pacbio_reads, clc_extract_consensus and clc_mapper.
• Added support for floating point values in the clc_sam_to_cas tool.

Bug fixes

• Fixed an issue in clc_mapper where a short deletion was preferred to a mismatch for equal scoring alignments
• Fixed an issue in clc_correct_pacbio_reads where it would fail when it encountered reads containing an ambiguous symbol. Such reads are now ignored.
• Fixed an issue in clc_correct_pacbio_reads that caused it to fail if input reads contained long homopolymeric regions.
• Fixed an issue in clc_sam_to_cas where the tool would have crashed when encountering an asterisk symbol indicating an alternate alignment for the same read in the input SAM/BAM file.

Changes

Support for the deprecated scarf file format has been removed.

Bugs

• Fixed a bug in ‘clc_mapper’ that led to some deletions being reported as multiple, separate deletions instead of a single, larger deletion, when affine gap costs were used.

Advanced notice

The tools ‘clc_mapper_legacy’ and ‘clc_remove_duplicates’ will be retired starting from Assembly Cell 6.0 onward. They will remain present on older release lines. If you are concerned about the proposed changes, please contact our Support team ts-bioinformatics@qiagen.com

Improvements

• Added a new parameter –normaltypelimit (-S) to ‘clc_mapper’ for specifying the read length to be applied when deciding when to switch to long read mapping. The default value is 500 (bp), which guarantees the same default behavior as the previous version (Assembly Cell 5.0.5).
• Added a new parameter –mappermemory (-M) to ‘clc_mapper’ that allows the amount of memory (MB) to be used during mapping to be restricted. This parameter needs to be set when –normaltypelimit (-S) is used, since the memory usage increases with the squared read length, which can potentially cause the system to run out of memory. The -M parameter overrides the –cpus parameter to ensure a lower memory usage.

Bugs

• Fixed a very rare bug in ‘clc_mapper’, where an alignment with a leading unaligned end could result in a wrong score
• Fixed  a bug where ‘clc_correct_pacbio_reads’ would segfault if not given an output file
• Fixed a bug where the argument –interleave was not accepted by the tool ‘clc_quality_trim’

Changes

• 32 bit versions of the QIAGEN CLC Assembly Cell are no longer distributed.

Advanced notice

The tools ‘clc_mapper_legacy’ and ‘clc_remove_duplicates’ will be retired starting from Assembly Cell 6.0 onward. They will remain present on older release lines. If you are concerned about the proposed changes, please contact our Support team ts-bioinformatics@qiagen.com

Bugs

• Fixed a very rare issue, where a read could be aligned incorrectly and would be reported as not mapped.

Advanced notice

From the next feature release of QIAGEN CLC Assembly Cell and forward, 32 bit versions of the QIAGEN CLC Assembly Cell will no longer be distributed.

Bugs

• Fixed an issue in the clc_mapper tool where length and similarity fraction cut-offs were ignored for reads where the best read alignment was a stretch of exact matches, with no extension from that seed possible, and where the read was longer than 500bp.
• Fixed an issue in the clc_mapper tool that caused it to crash if a sequence with length 0 was included in the reference set.

Advanced notice

From the next feature release of QIAGEN CLC Assembly Cell and forward, 32 bit versions of the QIAGEN CLC Assembly Cell will no longer be distributed.

Bugs

• Fixed a bug in the read mapper causing it to fail on CPUs with no support for SSSE3.
• Fixed a bug that sometimes made the index building phase of the read mapper crash or freeze when run with many references and many threads.

Advanced Notice

From the next feature release of QIAGEN CLC Assembly Cell and forward, 32 bit versions of the QIAGEN CLC Assembly Cell will no longer be distributed.

Improvements

• The maximum memory requirement for read mapper index building has been decreased from three quarters to half of the physical memory.
• Minor improvements to alignment quality of read mapping.

Bug fixes

• Fixed an issue where the distance between paired reads could be miscalculated after read mapping if one of the two reads had an unaligned end. This lead to some pairs erroneously being marked as broken.
• Read mapper index building now takes account of the –cpus parameter.
• Cas file output from read mapping now reports sequence circularity correctly.
• Default values in the usage information for the clc_assembler_long parameters minimum contig length, minimum word coverage and minimum anchor length have been corrected.

Changes

• Random sampling between equally scoring alignments produced by read mapping has been changed to better mimic the behaviour in Assembly Cell 4.4.

Advanced Notice

From the next feature release of QIAGEN CLC Assembly Cell and forward, 32 bit versions of the QIAGEN CLC Assembly Cell will no longer be distributed.

Bug fixes

• Fixed a bug in the clc_correct_pacbio_reads tool that caused the tool to crash when run on a machine without SSE3 support.
• Fixed a very rare bug in the clc_mapper tool that sometimes caused mapping positions to be reported multiple times.
• Fixed a bug in the clc_cell_licutil licensing tool that caused the evaluation licenses generated by the tool to be invalid for the current version of QIAGEN CLC Assembly Cell.

Advanced Notice

From the next feature release of QIAGEN CLC Assembly Cell and forward, 32 bit versions of the QIAGEN CLC Assembly Cell will no longer be distributed.

New tools and features

• clc_correct_pacbio_reads: correct PacBio reads as a preprocessing step before assembling them.
• clc_assembler_long: de novo assemble PacBio reads that have already been corrected using for example clc_correct_pacbio_reads.

Improvements

• clc_mapper: This update
  • Will greatly improve read mapping quality with long reads (PacBio or Oxford Nanopore).
  • Will return the same read mappings as before with short reads (Illumina and ion torrent) except from a few minor bugs that have been fixed.
  • Will improve read mappings with longer reads (454) slightly.
  • Allows the user to set the match score with the new –matchscore parameter.

Retired tools

• clc_find_variants: has been deprecated and replaced by clc_extract_consensus since QIAGEN CLC Assembly Cell version 4.2.

Bug fixes

• Fixed a bug that caused adapters to only be searched in the forward direction when specified for only odd-numbered or only even-numbered reads.
• Fixed a bug that messed up quality scores when reading from two interleaving paired files.
• Two consecutive mismatches in the MD field of SAM/BAM files produced by clc_cas_to_sam are now separated by a “0”, which they should be according to the SAM format specification.
• Fixed a bug in clc_cas_to_sam that caused soft clipped ends to be included in the count for the NM field in SAM/BAM files produced by the tool.
• Fixed a bug that caused the license check to fail on Mac OS X 10.11.3 and higher.
• Fixed a bug that made the clc_assembler tool fail when the “–interleave” option was used.
• Fixed a bug in the clc_extract_consensus that caused the tool to not recognize the option “-r unknown”.

Advanced Notice

From the next feature release of QIAGEN CLC Assembly Cell and forward, 32 bit versions of the QIAGEN CLC Assembly Cell will no longer be distributed.