3DNA Forum

Questions and answers => General discussions (Q&As) => Topic started by: oskolkov on February 18, 2013, 06:27:50 am

Netiquette · Download · News · Gallery · Homepage · DSSR Manual · G-quadruplexes · DSSR-Jmol · DSSR-PyMOL · DSSR Licensing · Video Overview· RNA Covers

Title: Building chromosome with fiber
Post by: oskolkov on February 18, 2013, 06:27:50 am
Dear all,

I would like to ask if anyone tried to build a structure of whole chromosome with fiber? I tried and got a message "allocation failure 1 in cmatrix()", so I suspect that the sequence was too long. Can anyone recommend me how to build a chromosome with fiber or with any other program? Thank you!

Cheers!
Nikolay.
Title: Re: Building chromosome with fiber
Post by: xiangjun on February 18, 2013, 10:18:38 am
The error message you reported is related to memory allocation. It is likely that your computer is not up to the task. What are the settings of your computer? How much RAM do you have?

How did you run the fiber program specifically? How many bps does the chromosome have?

If you could provide such details, others may help you identify where the problem is, and possibly come up with a solution.

Xiang-Jun
Title: Re: Building chromosome with fiber
Post by: oskolkov on February 18, 2013, 10:29:58 am
Hi Xiang-Jun,

thank you for your reply, I have 4 Gb RAM and I tried to build human 21 chromosome which contains around 48 000 000 bps. I monitored my memory usage and did not notice that fiber used a lot of memory. Do you think I should try to build it on a computer with more memory?

Cheers!
Nikolay.
Title: Re: Building chromosome with fiber
Post by: xiangjun on February 18, 2013, 12:07:40 pm
Hi Nikolay,

Thanks for providing further info. I tried to build a fiber-based structure with 48M bps on my new iMac with 32GM RAM. The fiber program did not exist with the memory allocation error as you saw, but it seems never to finish!

So fiber appears not up to the job, even though in principle it should. Note also that with 48M bps, the PDB format is certainly out of the question. If you really want to go for it, you may try smaller fragments (with the -xml option for the PDBML format) that can be built with fiber, and then assemble them into the whole 48M bps chromosome.

Xiang-Jun

Created and maintained by Dr. Xiang-Jun Lu [律祥俊] (xiangjun@x3dna.org)
The Bussemaker Laboratory at the Department of Biological Sciences, Columbia University.