SNV
Tools | Pros | Cons | Used in Logan |
Mutect2 | Part of GATK best practices | | x |
Strelka | Fast | Paired only | x |
Muse | Fast | Paired only, can't be parallelized | x |
Lofreq | Low frequency variants | Slow,Paired only | x |
Vardict | Fast | Lower accuracy | x |
Varscan | Fast | Lower accuracy | x |
Octopus | Accurate | Slow,High memory | x |
Deepsomatic | Relatively fast | Trained on human data | x |
Structural Variants
Tools | Pros | Cons | Approach | Used in Logan |
Manta | Accurate, fast | | graph-based | x |
SVABA | Deletion detection | | local assembly+ multiple alignment | x |
GRIDSS | Provides blacklist | Slow, part of HMFtools pipeline | Break end assembly (discordant +split) | x |
Manta, GridSS, and SvABA are based on read-pairs, split-reads, and local-assemblies.
References Joe et al
Copy Number
Tools | Pros | Cons | Used in Logan |
Purple | Complete workflow | Doesn't support mm10, requires SV,SNV calls as well | x |
Sequenza | Purity/Ploidy | | x |
FREEC | Fast | No Purity/Ploidy Estimatation | x |
ASCAT | Fast, Purity/Ploidy | | x |
CNVkit | Fast | No Purity/Ploidy Estimatation | x |
PureCN | Tumor only | Needs Panel of Normals on Sequencing | |
Germline
Tools | Pros | Cons | Used in Logan |
Deepvariant | Fast, most accurate | Model trained on human genomes (May not support mm10) | x |