Non-uniq names in FASTA

[ercfrtz]

I have a fasta file I created from the bovine gene information and I ran a uniq -d command on the names to make sure I didn't have any name duplicated. But when I use it as a reference and align reads to it and then try to run those reads through picard. Picard tells me I have duplicate sam sequences.

Does anyone know of a simple solution to this or have a way to identify those troubling sequences that appear to have the same name, even though the uniq command won't identify them?

[pmiguel]

On the sorted names?
Try:

grep ">" file.fasta | sort |uniq -d

If you get nothing, try:

perl -lane '$a{$F[0]}++;END{for (keys %a){print if $a{$_}>1;}}'

If still nothing, something else if probably going on. Not clear what you mean by:

"Picard tells me I have duplicate sam sequences"

Is "sam" a typo or you are working with a SAM file?

--
Phillip

[sklages]

I have a fasta file I created from the bovine gene information and I ran a uniq -d command on the names to make sure I didn't have any name duplicated. But when I use it as a reference and align reads to it and then try to run those reads through picard. Picard tells me I have duplicate sam sequences.

Does anyone know of a simple solution to this or have a way to identify those troubling sequences that appear to have the same name, even though the uniq command won't identify them?

How have you applied your 'uniq' command? Keep in mind that the fasta name is everything up to the first whitespace in the definition line. So something like

>bumblebee is great

and

>bumblebee is yellow

is for most programs the same name. Uniq'ing the definition lines is not enough.

Sven

[ercfrtz]

It looks like my file wasn't sorted and I thought it was, so uniq wasn't catching certain things. Using sort into uniq fixed the issue. Thanks for the replies.