Version 1.0.8 stable

• Better phage blast parallelization
• A place for predefined settings
• Detecting "cliffs" in read coverage. Places where recombination likely occurred and where base calls have less reliable
• Added clinical isolates to the Lmono dataset

Stable

• new Lyve-SET.pm module
• better diagnosing script for MSAs
• percent identity threshold for launch_smalt

v1.0.4

• --clean totally removed in preparation for when we bring in quake or another cleaning tool
• --fast: downsamples, samples variant sites, does not find phages, does not use a sample assembly when testing, snap
• Fixed launch_snap a bit more, for stability and for unambiguous matches
• Better documentation including tips and tricks, FAQ, and a different installation document
• --singleend for when you have mixed datasets of PE and SE reads

Ability to mask regions; fixed several bugs; more transparency in output and in documentation

• able to restart set_test instead of having to remove the directory before restarting it
• Fixed vcfMerge.sh bug - looking for the wrong executable
• All MSA and SNP analysis is done on the matrix/bcftools level
• Masking the reference genome is possible through BED files.
  • There is a phage finder using the phast database, and it puts in masked regions
  • masking is done on the read-mapping level and so some reads will traverse the indicated 'bad regions'
• simulating reads is faster just by reducing the number of individual jobs
• Simulated reads are 100x instead of a fixed number. The number of reads is now more appropriate.
• fq and fq.gz extensions are now recognized
• Removed freebayes - it never calls invariant bases and so it is fundamentally wrong for this pipeline
• VarScan and future snp calling modules will be expected to have filtered output before the rest of the pipeline has seen it. See launch_varscan.pl for how it's done.
• More documentation like in FAQ.md
• Request in README.md for users to join Google Groups

Feels pretty stable now; installation is easier

• more prerequisites installed with make such as raxml v8
• wgsim parameters updated
• make clean is more comprehensive
• Lmono dataset is complete finally (but maybe some clinicals should be added)
• refactored mvcf to aln and now have ambiguities allowed by default
• Bugs found and fixed

v0.9.8: Like v0.9.6 but better installation

updated version to 0.9.8

v0.9.6

Changes:

• Phage finder (phast)
• Scripts to test Lyve-SET better (set_test.pl)
• Script to download test data
• Better documentation
• More stable Schedule::SGELK
• Window size for SMALT is 13, step size is 2 (before it was 5 and 3 respectively, making it needlessly slow)
• Varscan more stable

Multisample VCFs; no read cleaning; test data available

• Removed cleaning options -- users must clean reads themselves
• Varscan has been added and is the default SNP caller
• Using pooled VCF multisample file for the basis of the MSA/phylogeny
• Needs newest samtools/bcftools
• Can download test sets
• Underscores allowed in the ref genome

Uses hqSNP matrix; many bugs fixed

v0.9.2

version 0.9.2

Uses hqSNP matrix

Lyve-SET now is forced to go through the hqSNP matrix before creating a MSA. This creates a "paper trail" of where SNPs are located and so that it is easier to validate and trace what happened in each run. Any ambiguities or uninformative sites or clustered SNPs are removed in the matrix stage so that each run of Lyve-SET can be traced better. There is also a log directory where all qsubbed scripts, etc document their output.