Hi all, 

I'm trying to write an FCS file using the write.FCS function and am noticing some (perhaps unexpected) behavior that leads to data truncation. 

Specifically, if I create a flowFrame object using a matrix of values, do not set the channel ranges explicitly, and then write the flowFrame to disk using the write.FCS file, the channel ranges are by default set to 1024. Thus, when I have values outside this range, write the values to a file using write.FCS,  and read that same FCS file back, the values outside of that range have been truncated to channel min/max values.

The line that is performing this behavior is here:

https://github.com/Bioconductor-mirror/flowCore/blob/e780de2897dc00fd5df6af59f03365ed5bd80106/R/IO.R#L1162

Is this the expected default behavior for when a range of data is not explicitly provided for a channel? I guess I would have expected that the range be automatically set to the observed range of data, rather than a default of 1024.

Thanks for any insight / help.

Cheers, 

-Rob

I'd say you should probably set 'PnR' (i.e. the instrument range) keyword for your flowFrame in the first place because it is required based on FCS3.1 http://isac-net.org/PDFS/90/9090600d-19be-460d-83fc-f8a8b004e0f9.pdf , section 3.2.18.

Alternatively, you can set 'truncate_max_range' to FALSE when read.FCS.

Thanks for your response Mike.

I totally agree that it should be set (as it is a required parameter). 

I guess what I'm interested in hearing is if you think that the write.FCS function should (in the absence of that keyword explicitly being set by the user) set the PnR values to the observed range of the data by default rather than '1024'. The latter seems a bit arbitrary to me (unless that too, is part of the 3.1 spec).