Take the 2-minute tour ×
Biology Stack Exchange is a question and answer site for biology researchers, academics, and students. It's 100% free, no registration required.

I'm working on a project where I am taking DNA sequences and translating the codons into musical notes. I have some good ideas of how to do this, I'm just not sure what sequences to work with. My case study and a lot of my initial research showed that anything larger than a protein is too much to work with for now.

So my question. Where can I find protein DNA sequences in some standard format. I've looked at NCBI but I have no idea what I'm looking at or if I'm downloading the right stuff. Is there a link that I can go to that will have a listing of a bunch of difference protein sequences that I can download?

share|improve this question
add comment

3 Answers

up vote 8 down vote accepted

GenBank and RefSeq have a huge collection of DNA sequences that can be downloaded in Fasta format. GenBank makes it very easy to search for sequences (for example, using the organism name) but will often have redundant data and varied types (everything from whole chromosome sequences to mRNAs to gene sequences to ESTs). RefSeq has much cleaner, non-redundant data but might take a little more effort for you to find the sequences you want. Depending on whether this has any effect on your project, you might want to consider that. On the RefSeq page, under the FTP menu you can click Genomes to see which genomes have RefSeq data available. The Fasta sequences for a particular genome will be stored in a dedicated directory, possibly subdivided into further directories by chromosome. What you want are the Fasta files (ending in .fa, .fasta, .ffn, or something like that), keeping in mind they may be compressed (.fa.gz, fasta.gz, etc).

Alternatively, if you already have a specific organism in mind, then you could do a Google search to see if the genome for that organism has been sequenced. If so, there is almost always a dedicated website where you can download DNA sequences for that genome.

share|improve this answer
    
I've been to those sites, I just don't know where to go to get a list of sequences. I want a wide variety of sequences so a list or the root of a directory would help the most. I'm not looking for a specific sequences. I just need a lot of them. –  justausr Apr 20 '12 at 14:57
    
In GenBank for example, if you search for Vitis vinifera, you get 131k reults. If you click the "Send to" button, select file, and then Fasta format, it will allow you to download a Fasta file with the sequences. If that's not what you're looking for, you're going to have to provide a little bit more detail before we can help you. –  Daniel Standage Apr 20 '12 at 15:01
    
I don't know a lot about biology. I'm researching different ways to encode data to help with perception of the data. So I know alot about information, just not DNA. I need a bunch of sequences to use as tests on my DNA to music tool. –  justausr Apr 20 '12 at 15:03
    
Is 131k sequences enough? –  Daniel Standage Apr 20 '12 at 15:07
1  
Ok this is a good answer I did what you said and have exactly what I wanted, actually a little better now that I see I can search specific organisms. –  justausr Apr 20 '12 at 15:14
add comment

This is more of a comment, but too long to put in a comment box, so I'm putting it here.

This is a fun idea you're doing. I have a half baked idea (assuming you're looking for input) if you want to explore it further -- or not ... by the time I finish writing this, I might realize it's too silly, but still ... let's see.

It might be fun to take the sequence of two organisms, let's say mouse and human and align certain regions to each other -- imagine this is like playing a piano where the "left hand" might be the mouse sequence, and the "right hand" is human.

So, say you take a gene that are shared in both, like CCND1. You can align them against each other and you'll find large portions of the sequences are common (with some mismatches, obviously). In these regions, the left and right hands are playing together (different octaves.

You'll also find gaps in the alignments where you'll have a stretch of "mouse only" or "human only" sequence, and in these regions the left or right hand play alone (solo).

For instance, say the two alignments look like this:

mouse: CGTGGGAGGCTCTTGAGCCTGGAAACACTATCGCAGTTTGTACGGAATGCACTTGTTCTTTACAAAAGG
human: CTTGGGCGACA---GAGC---GAGACTTTGTCTCAAAAAAGAAG--------------------AAAAG

In this case, you see stretches of the alignments where the mouse (left hand) will be playing a solo, and other times the two hands play in "harmony."

share|improve this answer
    
The tool I'm making can be used to do this without any modification. Cool idea though. I may make a blog for it once I finish the hard parts and post it to this QA just for reference. –  justausr Apr 20 '12 at 19:12
    
Here's the tool dnasongs.brandonhamric.com. The alignment part is left out because of some bugs in playing multiple parts at once, but oh well. –  justausr Apr 22 '12 at 6:38
add comment

A standard format would be the FASTA format.

If you have a few proteins of interest that you would like to look at you can simply select 'Nucleotide' at the top of the NCBI page, enter your protein's name and then press the FASTA button below each result you are interested in.

If you just need a large dataset of DNA sequences that code for proteins you could for example use the 'nt' blast database. (I believe this is only protein sequences not entirely sure though, I thought it would be easier to get)

share|improve this answer
    
I've seen that format too, is there a place I can go and download a lot of sequences in that format? –  justausr Apr 20 '12 at 14:42
    
good answer. I'll download the 10 gb file and get to work haha. Thanks –  justausr Apr 20 '12 at 15:00
4  
Since I can only add comments here: I agree with Daniel Standage that Genbank is the more suitable ressource but how do you ensure there that it's only protein coding DNA? You could limit your search for mRNA on the right which should throw out some non-protein results however then you have mRNA sequences which of course differ from the DNA sequence they originate from. If that is not an issue I'd say go for that. –  mivilar Apr 20 '12 at 15:23
    
This suggestion helped a lot –  justausr Apr 20 '12 at 19:14
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.