Skip to content

Latest commit

 

History

History
42 lines (29 loc) · 3.1 KB

history.md

File metadata and controls

42 lines (29 loc) · 3.1 KB
Raw

Latest version: v.2.3.0

  • add --fast option for clust-mst to use the more efficient Kssd sketch strategy when computing the all-vs-all genome distances.
    • The --fast option can work together with --append, --presketched, and --premsted options.
    • the --drlevel is used for setting the dimention reduction level for Kssd sketches. Default value is 3, which is corresponding to a dimention reduction of $1 / 2^{(4*3)} = 1/4096$.

v.2.2.1

  • add --newick-tree option to output the Newick tree format for clust-mst.

v.2.2.0

  • support incrementally clustering by option --append accompanied with --presketched or --premsted options.

Note:

  • When considering the clustering of the genome set A+B using a pre-generated sketch A_sketch and an appending genome set B, it is important to note that the sketch parameter for the pre-generated sketch A_sketch and the appending set B may differ from that of the whole genome set A+B. However, the impact of changes in the genome lengths of set B on the automatically generated parameters will be minimal if they are not significant.

    • This is because the sketch parameters, including the $k$-mer size, sketch size, and containment compress ratio, for the appending genome set B are the same as those of the pre-generated sketch A_sketch. Additionally, the automatic parameter generation method, which is carried out using the tune_parameters() function, depends on whole genome information such as minimum, maximum, and mean genome length. Therefore, the changes in the genome lengths of the appending set B are unlikely to have a significant effect on the automatically generated parameters if they are not substantial.

  • In the context of genome clustering, the sketches are sorted by unstable sort in a decreasing order of their genome length. Consequently, the order of sketches may undergo slight changes if there are genomes with identical lengths. However, this does not significantly affect the outcome of the clustering process.

v.2.1.0

  • change the parameter parsing by CLI11.
  • save the intermediate files (sketch, mst files) in binary format.
  • abrogate the -f option for loading pre-generated sketch or MST file, replaced by --presketched and --premsted option.

More details by clust-mst --help or clust-greedy --help.

v.2.0.3

  • add the parameter -m to set the minimum genome length (minLen), genomes with lengths less than minLen will be ignored.

v.2.0.2

  • update the calSize of gz files for automatically generating $k$-mer size .

v.2.0.1

  • Update the latest version of robin-hood-hashing to solve the compile error with g++ 12.0+.

v.2.0.0

  • Add the clust-greedy module for greedy incremental clustering.
  • Last MST-based clustering module is clust-mst module.

v.1.0.0

  • First version of RabbitTClust, large-scaled genome clustering tool based on sketch technique and Minimum Spanning Tree (MST).