Play it again Sam, but double the tempo

• Now runs 2 to 3 times faster due to new threading support in markdup and fixmate 😍
• All required linux binaries are now bundled to simplify container building
• Reinstated TRAVIS testing
• Improved CPU spreading logic for main alignment pipeline stages
• Variants near contig ends will be less likely to be missed due to improved samclip
• Bundled tools now fully python3 compatible

Keeping up with Sam

• fixed version parsing bug with samtools 1.10
• fix minor freebayes parameter bug when depth was close to --mincov
• explicitly set --sample in bcftools consensus to avoid --sample option not given warning
• increase minimum version of tools needed (see snippy --check)

Grey genotypes

Fixed a bug where variants of low genotype quality (QUAL) were being filtered out (by --minqual) but not being masked from the alignment. This meant these sites would assume the default reference allelle, but often were low QUAL due to strand bias and other problems.

It only really affects people using snippy-core and will not change the SNPs you would call, but will affect the final core genome you get from a set of isolates. This is a significant change in some cases, hence the bump to 4.1.0

Thank you to @kmeaton for alerting me to this problem.

Important bug fixes

• fix frame for snpEff
• un-break VCF output for extracting subs
• don't print null vcf tags
• handle prefixes that aren't snps in core
• use final dir in path to folder as label in core
• remove useless scripts in bin

Major overhaul and bug fixes

bug fixes

• yes, plenty. please upgrade!

snippy

• diploid mode to ensure mixed/het sites not used
• better BAM using fixmate and markdup
• filter soft clipped reads with samclip to avoid false positives
• better freebayes parameters to handle low complexity regions
• more bcftools and vt, less perl scripts
• parallel stack reports for --report (currently too slow for lots of SNPs)
• consistent --version reporting across tools
• cleaner categorised --help output
• new --targets option to focus on BED regions
• new --check option to validate installation
• new --minqual for VCF quality filtering

snippy-core

• complete rewrite
• --noref no longer supported
• ability to mask with a --mask BED file eg. phage regions from PHASTER
• more use of alphabet to denote SNPs and het sites

general

• better TRAVIS tests
• removed unused output files
• updated bundled binaries
• new dependencies vt and bcftools

Bigger than π

• Better docs
• Many bug fixes
• New snippy-core --ref option as to not require auto ref discovery in first snippy folder
• Improved end coverage for --ctgs option
• Newer static binaries where possible
• Better logging
• TRAVIS-CI support

Static binaries

For Galaxy packaging mainly.

Contigs vs Contigs

• You can now find SNPs between a reference and a set of contigs with the --ctgs option
• Updated binaries including samtools 1.3
• Wanted to avoid version 2.10 due to problems comparing semantic versions

Attack of the feature table

• major bug fix / work-around to handle codon_start problems in snpEff
• new --unmapped option to save unaligned reads to .fq.gz
• fixed column headers in .tab files
• bug fix for corner cases divide by zero

Halloween Edition!

I'm sorry for any indels you missed with old releases, it seems Freebayes and I don't get along!

You really need to uprade to this version!