FCS Files: JMP opens them, but where does the header information go?

[#1]

I am helping a colleague with some work with FCS files. JMP opens the data easy enough, which is great, but I am looking to extract the header information, and I can't figure out an easy way - especially if I wanted to do this for a bunch of files.

I don't get any options like I do for CSV files, so the only solution I have found is to change the extension to .txt and open the file with text preferences. That works, but would seem troublesome if I need to do this to all files prior to opening.

1) is there a way to open a single FCS file to see the header w/o changing the extension
2) can you use a script and force-use text preferences even if the extension isn't a standard text extension.

Thanks!

[#2]

When JMP opens an FCS file it places the header information as table variables.

You can interrogate the table to get the list of names (dt<<getTableVariableNames) and you can retrieve the values (

dt<<getTableVariableNames). 

[#3]

@David_Burnham, That would be awesome! ... but I am not seeing that. I am on JMP 15.2.1 on Windows 10.

[#4]

One option you might want to try is using the function LoadTextFile, this will at least give you visibility to the header data (although can be slow because FCS files contain large blocks of binary data):

fcsFile = "c:\my-fcs-file.fcs";
str = loadtextfile(fcsFile);
// extract header
pat = "FCS" + PatArb()>>hdr + "$BEGINANALYSIS";
isMatch = PatMatch(str,pat);
if (isMatch,
	show(hdr)
,
	show("oops")
);

[#5]

I will give that a try ... Opening the file in a text editor the header has a lot of information (apparently such that JMP isn't pulling it out as header information).

Maybe this will give me a hand understanding why...

Thanks!

[#6 · Accepted solution]

I took the code a bit further to unpick some of the fields in the header

fcsFile = "my-fcs.fcs";
str = loadtextfile(fcsFile);
// extract header
pat = "FCS" + PatArb()>>hdr + "$BEGINANALYSIS";
isMatch = PatMatch(str,pat);
if (!isMatch,
	throw("oops")
);
// extract selected header info
system = ""; 	// computer system
pat = "$SYS" + "|" + PatArb()>>system + "|";
PatMatch(hdr,pat);
cytometer = ""; // type of cytometer
pat = "$CYT" + "|" + PatArb()>>cytometer + "|";
PatMatch(hdr,pat);
startTime = ""; // start time
pat = "$BTIM" + "|" + PatArb()>>startTime + "|";
PatMatch(hdr,pat);
endTime = ""; // end time
pat = "$ETIM" + "|" + PatArb()>>endTime + "|";
PatMatch(hdr,pat);

show(system,cytometer,startTime,endTime);

For my test data file this gives

system = "Summit V5.2.0.7477 / Windows NT Version 5.1 / Intel Processor";
cytometer = "MoFlo XDP";
startTime = "12:34:07";
endTime = "12:38:15";

I used some field identifiers based on this documentation

https://www.bioconductor.org/packages/release/bioc/vignettes/flowCore/inst/doc/fcs3.html#3.1

[#7]

Thanks! I will need to touch base with support to see why this file doesn't have the header info in the table vars, but I can use the script you sent to at least dig into the files. Even with the overhead of loading an entire file, it beats the alternatives

-Nate

[#8]

Yes I am curious why you are not seeing the header information.  I would guess either it is the version of the FCS files (I was testing with FCS3.0), or JMP is selective in the header information that it looks for.

[#9]

I'm pretty sure JMP has a list of standard keys it recognizes and may ignore other (vendor specific?) keys.

There is also information in each column's properties, maybe for interpreting ranges.

@Anonymous @briancorcoran