Sailfish equivalent to bowtie options -v and -all

[Sebastian Boegel]

Hi,

i am new to sailfish. I was using bowtie all the time. Is there any chance that i can say sailfish, that it should report all possible mappings (in bowtie -all) and that i can set the number of allwoed mismatches?

Thank you in advance

[Rob]

Hi Sebastian,

  Since sailfish is based purely on k-mers, there is no notion of an allowed number of mismatches.  A k-mer from a read either matches exactly to some place in the reference or it does not.  Every k-mer from the read is considered, so there may very well be mismatches (or gaps) in the read while it's still producing matching k-mers.  Regarding "all" mapping, sailfish, by default, considers all possible mapping locations of a k-mer, and all possible locations are considered during quantification as well --- so you need not do anything extra to enable that behavior.

  As I've been saying to most of the users posting here, I also encourage you to test out our new tool salmon (source repo, latest release) which has two modes of operation.  The "alignment-free" (or, perhaps more appropriately, "lightweight-alignment") mode allows you to specify how much of a read should be covered by exact matches (MEMs) before that read should be considered to align with a location in the reference.  It will also allow you to specify the maximum number of mapping locations that should be considered.  Salmon also has an alignment-based mode where you can feed it a bam/sam file generated by e.g. Bowtie.  This would allow you to examine how the different mapping strategies affect quantification estimates while holding fixed the actual inference procedure.

Best,
Rob