Clinical PathoScope pipeline. A computational subtraction method using varying sequence read lengths and ambiguous read reassignment. Unassembled sequencing reads are aligned against a target library containing reference sequences of the intended target(s) of identification (e.g. viruses). Reads aligned to the target library are then aligned to a host library. Any reads aligned to the host sequences are removed from further analysis. Next, reads are aligned against a library of known non-target sequences. Unaligned reads are then mapped back to the target library, allowing up to k alignments per read (e.g. k = 10). These alignments are subsequently passed to an expectation maximization algorithm in which ambiguous alignments are reassigned to their most probable genome of origin. Upon reassignment, a report detailing the pathogens identified and their relative abundances is produced.