Flye 2.7.1

• Fixes very long GFA generation time for some large assemblies (no other changes)

Flye v2.7

• Better assemblies of real (and comlpex) metagenomes
• New option to retain alternative haplotypes, rather than collapsing them (--keep-haplotypes)
• PacBio HiFi mode
• Using Bam instead of Sam to reduce storage requirements and IO load
• Improved human assemblies
• Annotation of alternative contigs
• Better polishing quality for the newest ONT datasets
• Trestle module is disabled by default (use --trestle to enable)
• Many big fixes and improvements

Flye 2.6

• This release introduces Python 3 support (no other functional changes)

Flye 2.5

• Better ONT polishing for the latest basecallers (Guppy/flipflop)
• Improved consensus quality of repetitive regions
• More contigouous assemblies of real metagenomes
• Improvements for human genome assemblies
• Various bugfixes and performance optimizations

Flye 2.4.2

• Improvements in k-mer selection and tip clipping for metagenome assemblies
• Better memory managment during consensus/polishing
• Some bugfixes

Flye 2.4.1

• Speed and stability improvements for large datasets
• New option --polish-target to run Flye polisher on the target sequence

Flye 2.4

• Metagenome assembly support fully integrated (--meta option)
• New Trestle module for resolving simple unbridged repeats
• New --plasmids option that recovers short unassmbled plasmids

Flye 2.3.7

• Improvements in repeat edges detection
• More precise read mapping - more contiguous assemblies for some datasets
• Memory and performance optimizations for high-coverage datasets
• More accurate repeat graphs for complex datasets

Flye 2.3.6

• Memory consumption for large genome assemblies reduced by ~30%
• It could be reduced even further by using the new option --asm-coverage,
  which specifies a subset of reads for initial contig assembly
• Better repeat graph representation for complex genomes
• Various bugfixes and stability improvements

Flye 2.3.5

• New read mapping implementation (based on solid k-mers) with improved specificity
• Better support of corrected reads input
• Minimum read overlap is now selected within a wider range for better support of datasets with shorter read length
• Assembly of large (human size) genomes is now faster
• Various bugfixes and stability improvements