MUMmer VS blastp

[anyone1985]

I just finish a bacteria genome, then compare two genomes of the same genus. I　hesitate between the MUMmer and blastp. Please give me some suggestions. And also, is there any software to get a dot plot from the result of blastp?

[bio-x]

yes, you can try BSR, it is based on blastp.

http://www.microbialgenomics.org/BSR/

good luck!

[anyone1985]

Thank you for bio-x's kind help. I think accurately I'd like to get a dot plot of two genomes's ortholog plot by reciprocal best-hit pair from the result of pair-wise BLASTP. I guess there must be some software which is suitable to do this. Please give me some suggestions.

[Torst]

I think accurately I'd like to get a dot plot of two genomes's ortholog plot by reciprocal best-hit pair from the result of pair-wise BLASTP.

"promer" from the Mummer suite will effectively do want you want. Although it doesn't examine "genes" it matches translated regions, which if there is a good match, are likely to be orthologs (or paralogs) anyway.

[anyone1985]

Yes, promer is pretty good. However, I'd like to draw the graph exactly. The promer can not give a clear position to identify the gene. Many literatures have figures as below. I am intrested in which software can draw such figure.

"promer" from the Mummer suite will effectively do want you want. Although it doesn't examine "genes" it matches translated regions, which if there is a good match, are likely to be orthologs (or paralogs) anyway.

[rahul.m.dhodapkar]

does anyone know how to fix the following MUMmer error?

where textlen is the same as the length of build36 (hg18)

./mummer: suffix tree construction failed: textlen=3080436075 larger than maximal textlen=536870908

[robs]

Try:
http://seqanswers.com/wiki/Talk:MUMmer

The corresponding paper can be found here:

http://bioinformatics.oxfordjournals.org/cgi/content/full/25/13/1609

[rahul.m.dhodapkar]

@robs

thanks, but the link says:

[Thanks for the hint about the recent "MEMs" implementation that "can serve as a drop-in replacement for the MEMs algorithm in MUMmer 3". Seems to be working perfectly where MUMmer crashed :-) --Dan 15:08, 5 January 2010 (UTC)]

what was the original hint? (sorry for any stupidity)

[robs]

Not sure what you mean with "original hint". There is only one hint and if you take a look at the paper you will see that the quote is actually from there.
The MEMs implementation has a mummer module that can be used instead of the MUMmer3 one. By doing this, mummer shouldn't crash anymore. More details are in the paper.

[rahul.m.dhodapkar]

@robs

thanks so much for your help.

[Will Nelson]

To solve the textlen problem, build a 64-bit version, and make sure
#define SIXTYFOURBITS is set in types.h.
(Grep for the one place this is used, to see where to set it).

[flxlex]

To solve the textlen problem, build a 64-bit version, and make sure
#define SIXTYFOURBITS is set in types.h.
(Grep for the one place this is used, to see where to set it).

First, thanks for, as far as I can tell, the only solution to the mummer textlen problem on the web...

Not being an hardcore programmer, but a biologist-come-bioinformatician (more like a user of programs), it took me some time to figure out how to implement it, though. What you meant was to add this line

#define SIXTYFOURBITS

to the src/kurtz/libbasedir/types.h file before the line where it says

#ifdef SIXTYFOURBITS

and then run

make install

again.

(I had begun to 'uncomment' the ifdef line, causing all kinds of errors - like I said, not a programmer)

Just clarifying for future generations...

[colindaven]

This thread also supplies another solution:

http://sourceforge.net/projects/mummer/forums/forum/451664/topic/2072496

"
If you recompile the package like so:

> make clean
> make CPPFLAGS="-O3 -DSIXTYFOURBITS"

you should have no more problems with space limitations.

Best,
-Adam
"

[cindy0805]

I want to compare to genomes with 87%similarity, I cant decide should I use Promer or NUCMER?