What to expect for a sample that may have a mixture of two or more strains

[ilyavs]

Hi, This is more of a general question. Suppose I have a fastq file of a sample that has two or more strains of the same species, a contaminated sample. What could I expect from the assignment pipeline? Can it be used to detect the contamination? Thanks!

[johnlees]

We usually remove these samples. My expectation would be that it either joins these two clusters together, or forms a totally new cluster – depending on the nature of the contamination and the specific model fit used. I would look at this with some data (real or simulated), and in particular look at the core/accessory distances estimated by sketchlib to see if they are obviously outlying in a detectable/repeatable way

[ilyavs]

Thank you