[CNVboy]

Thx Prakhar,

Sorry for my delayed response. The problem of viewing has been solved. Yeah, you are right, use ssh -X and it works! thx~

I'm actually now trying larger-scale data on cluster, and would definitely ask for your kind help if come across new problems.

best regard

[zhk]

Is there anyone know the format of cnv result? For example:
deletion chr10:6070441-6074520 4080 0.674906 0.00834134 72162.3
It is very easy to understand the first three column mean, but is there any can explain the rest of column mean?

[gprakhar]

Dear zhk,

I had asked the same question to the author (Alex Abyzov), here is the reply.

The output is as follows:
CNV_type, coordinates, size, normalized_RD, p-val1, p-val2.
p-val1 -- is from t-test
p-val2 -- is from probability of RD valued within the region to be in
the tails of gaussian distribution (kind a z-score).

Regards,
pg

[zhk]

Thx gprakhar,

Is that mean the larger of p-val1 and p-val2, the more confidence of the result?

Do you find that the value of p-val2 will be very large or small sometimes? How to explain this situation?

example:
deletion chr10:6070441-6074520 4080 0.674906 0.00834134 72162.3

Dear zhk,

I had asked the same question to the author (Alex Abyzov), here is the reply.



Regards,
pg

[gprakhar]

Hello,

I am not sure, have used the tool on simulated data only.
You can ask the author and please do let me know what he says.

Regards,
pg

[zhk]

Hello pg,
This is the replay from the author.

Those p-values are not very discriminate. The major cause of false
positives is read mismapping and sequencing biases.
Large events are generally more trustful

[elisadouzi]

Hi milesgr,
I'm in the cluster and can not get into the root. How can I install the CNVnator?

Thanks!

Never mind, I figured it out. If anyone needs help, please feel free to post here.

[CNVboy]

Hi milesgr,
I'm in the cluster and can not get into the root. How can I install the CNVnator?

Thanks!

What do you mean by "cannot get into root". Have you downloaded ROOT package?
You'd better download the newest version of ROOT.(Old version may have bugs)
Then install CNVnator.

btw, ROOT is large package, you may take some patience.

[elisadouzi]

Thanks! I got it.

What do you mean by "cannot get into root". Have you downloaded ROOT package?
You'd better download the newest version of ROOT.(Old version may have bugs)
Then install CNVnator.

btw, ROOT is large package, you may take some patience.

[gprakhar]

Regards,
--
pg

[djeffares]

Problems making CNVnator

Hi all,

I've got root installed, but now I'm having problems compiling CNVnator. When I type make, I get this error:

g++ -pthread -m64 -I/usr/local/root/bin/root/include -Isamtools -c cnvnator.cpp -o obj/cnvnator.o
cc1plus: error: /usr/local/root/bin/root/include: Not a directory
make: *** [obj/cnvnator.o] Error 1

Any help much apreciated.
Dan

[dmacmillan]

Try simply by typing this:

export ROOTSYS=[path to root installed directory]

replacing everything inside and including the square brackets by whatever directory you installed root to - e.g. /home/djaffares/root
then type 'make' and hit enter and it should work! Good luck Dan!

[djeffares]

CNVnator, still not compiling

Thanks for this, but still no luck. here is what I did:
which root
/usr/local/bin/root

export ROOTSYS=/usr/local/bin/root
cd software/CNVnator/src/
make

#And I still get this error.

g++ -pthread -m64 -I/usr/local/bin/root/include -Isamtools -c cnvnator.cpp -o obj/cnvnator.o
cc1plus: error: /usr/local/bin/root/include: Not a directory
make: *** [obj/cnvnator.o] Error 1

#thanks in advance for any ideas :-)

[dmacmillan]

You have to "cd" into CNVnator/src/samtools and then "make", so libbam.a is created.

Also remember to type

export LD_LIBRARY_PATH=${LD_LIBRARY_PATH}:${ROOTSYS}/lib

Then type

cd ..

to get back into your CNVnator/src directory and type

make

again

[dmacmillan]

Also does anyone know what bin_size refers to? There is a bin size as in the .BAM bin size, but there is also the genomic bin size...