In silico PCR

[rhinoceros]

Does anyone know any good in silico PCR algorithm, preferably for offline use. Input would be a multifasta file and forward and reverse primer sequences. Output would preferably be substrings with intact headers.

[id0]

How about the tool by Jim Kent that's used by UCSC?

http://genome.ucsc.edu/cgi-bin/hgPcr?command=start

In-Silico PCR was written by Jim Kent. Interactive use on this web server is free to all. Sources and executables to run batch jobs on your own server are available free for academic, personal, and non-profit purposes. Non-exclusive commercial licenses are also available. Contact Jim for details.

[rhinoceros]

Yeah, this was one of the algorithms a little bird called Google suggested. I just didn't feel like contacting anyone. I suppose this might end up being the best option though..

[winsettz]

There's an hgPcr in the lastest copy of jksrc.zip. I've yet to to look into the code yet to determine if it's the same as the hgPcr in silico PCR on the website.

Edit: Or you can go to

http://genome-source.cse.ucsc.edu/gitweb/?p=kent.git;a=tree;f=src/hg/hgPcr;h=a22ca33e4a7f8e37b0e48750539136d05191e433;hb=HEAD

[rhinoceros]

There's an hgPcr in the lastest copy of jksrc.zip. I've yet to to look into the code yet to determine if it's the same as the hgPcr in silico PCR on the website.

Edit: Or you can go to

http://genome-source.cse.ucsc.edu/gi...91e433;hb=HEAD

Yeah, thanks. I was just going through the zip file. Problem is though that it has a bunch of dependencies and among other things requires a running mysql server. This is all a bit extensive for wanting to do just one relatively simple thing..

[mcnelson.phd]

mothur has a function called pcr.seqs that supposedly does what you're looking for. http://www.mothur.org/wiki/Pcr.seqs

[id0]

Yeah, thanks. I was just going through the zip file. Problem is though that it has a bunch of dependencies and among other things requires a running mysql server. This is all a bit extensive for wanting to do just one relatively simple thing..

That was my experience as well. I ended up building a scraper for the UCSC PCR page. It took a few seconds per lookup, so doing hundreds of searches took a while, but not an unreasonable amount of time. Unfortunately, if you try to perform multiple queries from the same IP concurrently, the IP is automatically blocked for a day, so there is no way to parallelize that.

[rhinoceros]

mothur has a function called pcr.seqs that supposedly does what you're looking for. http://www.mothur.org/wiki/Pcr.seqs

Yeah, that ought to get the job done. Thanks

[ScottC]

You can also try ipcress

http://manpages.ubuntu.com/manpages/raring/man1/ipcress.1.html

Cheers,

Scott.

[rhinoceros]

That was my experience as well. I ended up building a scraper for the UCSC PCR page. It took a few seconds per lookup, so doing hundreds of searches took a while, but not an unreasonable amount of time. Unfortunately, if you try to perform multiple queries from the same IP concurrently, the IP is automatically blocked for a day, so there is no way to parallelize that.

I contacted the author and turns out that source and binaries for many platforms are available via http://users.soe.ucsc.edu/~kent/

The provided macIntel binaries (faToTwoBit, gfServer, gfPcr) work almost out of box on recent Mac OS X. I only needed to symlink /opt/X11/lib/libpng14.14.dylib to /usr/local/lib

[rhinoceros]

So, I ended up using pcr_seq from biopieces. The below code yields the results I was expecting:

read_fasta -i input.fasta | pcr_seq -f CCTACGGGRSGCAGCAG[1,0,0] -R ACCGCGGCKGCTGGC[1,0,0] | grab -e 'REC_TYPE eq PCR' | write_fasta -x > PCR_products.fasta

-f and -R of course depend on what primers you want to use. There are also -F and -r parameters for other primer orientations. Within closed brackets, the number of allowed mismatches/deletions/insertions are specified. I found this to be the least tedious way to get the wanted output, in this case from 50k full bacterial 16S sequences I got 47.6k 'correct' PCR products. With gfPcr there's a huge number of options which makes it rather difficult to obtain desired output.

On related note, I hadn't used biopieces before. I have to say, I'm positively surprised. There seems to be a ton of very useful scripts included. E.g. plot_lendist appears to be a very easy (little work) way to plot length distribution with high quality output, i.e. to check the output of previous code (except on bacterial subset of Silva SSURef_NR99_115):

read_fasta -i PCR_products.fasta | plot_lendist -xk SEQ_LEN -t post -o plot.ps -L

Gives:


..works pretty well