>jenkins at aidsun.nibsc.ac.uk wrote:
>> : I have copied from the states the Los-Alamos hiv dbase (embl format).
>> : Using fromembl (unix flavour, SGI machine) I broke the flat file into
> : individual protein sequences.
>> : The output files containid "$" in the name eg env$u445.embl. My flavour
> : of unix will not recognise this dollar symbol, so i can't do anything
> : with these sequences.
>> I would suppose that you have troubles with the $ because it is a special
> sign for almost all unix shells, designating a shell variable.
>> There are two ways to circumvent this.
> 1) You can force the shell to recognize the $ sign by putting a backslash
> in front of it, e.g. env\$u445.embl instead of env$u445.embl.
> 2) You can quote the name of the file in single quotes, e.g.
> 'env$u445.embl' instead of 'env$u445.embl'.
>> Hope that helps,
Alternatively, you might want to replace all the $s in the original flat file
with an underscore (_) and then run the fromembl program to convert. Warren
Gish (NCBI) gave me this tip when I experienced a similar kind of problem with
a GenBank flatfile (I had to remove carriage returns). Here are the steps:
1. sed -e 's/\$/_/' hiv.flat > hiv.sub
(This line will replace the $ with _ and route it to the file, hiv.sub. Make
sure that you type the substitution parameter EXACTLY as above.)
2. fromembl hiv.sub (etc)
bollingt at ugene1.abbott.com