Latent strain analysis partitions reads into sets likely belonging to the same genome, using the covariance of k-mers across many samples.

Documentation for LSA, including a "getting started" tutorial with accompanying test data, and step-by-step instructions for analyzing large collections, can be found at: http://latentstrainanalysis.readthedocs.org/.

The source code can be found on the github page.

REFERENCE
Cleary, B., Brito, I. L., Huang, K., Gevers, D., Shea, T., Young, S., and Alm, E. J. (2015). Detection of low-abundance bacterial strains in metagenomic datasets by eigengenome partitioning. Nature biotechnology, 33(10), 1053.