With the development of next-generation sequencing, efficient tools are needed to handle millions of sequences in reasonable amounts of time. Sumaclust is a program developed by the LECA. Sumaclust aims to cluster sequences in a way that is fast and exact at the same time. This tool has been developed to be adapted to the type of data generated by DNA metabarcoding, i.e. entirely sequenced, short markers. Sumaclust clusters sequences using the same clustering algorithm as UCLUST and CD-HIT. This algorithm is mainly useful to detect the 'erroneous' sequences created during amplification and sequencing protocols, deriving from 'true' sequences. Currently, Sumaclust is available as a program that you can download and install on Unix-like machines.
Version 1.0.34: Compilation with the libsuma library (see https://git.metabarcoding.org/obitools/sumalibs).
Version 1.0.31: Fixed a memory bug with similarity thresholds of 100%.
Version 1.0.20: The input can now be the standard input.
Version 1.0.10: Sumatra and Sumaclust have been split in two packages.
Download the archive on this page, then untar it, go into the newly created directory and compile:
tar –zxvf sumaclust_v[x.x.xx].tar.gz cd sumaclust_v[x.x.xx] make -C sumalibs install make install
You can compile Sumaclust with clang, which deactivates OpenMP, with:
See the user manual downloadable from this page and included in the archive for a complete documentation.
Sumatra compares sequences using the same alignment methods as Sumaclust.