Lab Meetings Feb 2012


Wed Feb 1, 2012
******************
NOTE: john 21-24 of Feb  EPSCoRE 
	Feb 20/21/22 off for MG and Wed:w


	I WILL NEED TRAJECTORIES FOR THE
		 NUCLEOSOMES THAT MATCH UP W/ THE CSU SEQUENCES	 3x w/ and w/out water
		 ACGT control 16x  w/out water
		 biologic 6x   w/out water
		 yeast as much as possible w/out water

	PICKUP security IDs from Arlene on 1st floor Ian and Jacob (Kim/John do you have one?)

Rajib:
	data analysis
		-evector plots and analysis
			-10ns time periods for positioned nucleosomes as a movie
			-last 10ns for each position for the chromosomes
			questions: 
				do the max or min evector values "cluster"  in any manner
				what is significance of shape of the lines
		-contrib of each of the 12 helical parms to the total norm for each vector as function of position
			-10ns time periods for positioned nucleosomes as a movie
			-last 10ns for each position for the chromosomes
			questions:
				do all 12 helical parameters contribute evenly in all cases?
				does this mean energy is equipartitioned?
		-symmetry analysis: plot of LHS magnitude vs. RHS magnitude for each evector vs vector number
			-10ns time periods for positioned nucleosomes as a movie
			-last 10ns for each position for the chromosomes
			questions: 
				how many evectors associated w/ Hi evals are asymmetric? How many w/ Lo evals?
				how does this number compare to the number of kinks observed?

		I need open office compatible slides that define
			overlap, similarity and symmetry via mathematical formulas... as in the manuscript
			I need GRAPHICS and raw data that I can use in my talk 
			i need  to have the following data: 
					nowat.dcd, sys[parm,crd], par.dat files, kink data, evector-eval analysis, all movies we have made
				for the followin simulations
				a) the ACGT control simulations
				b) the positioned nucleosomes
				c) as many of the yeast chromosomes as we have
					

John/Kim: note: kim may be out
	need  copies of the following

	xls spread sheet of sequences from FesCUE
		-what is biologic function of each (promoter, gene, etc...)
		-where is the nucleosome positioned on each one (from the lit.)

	xls spread sheet of how  6 positioned nucleosome sequences compare
		-what is biologic function of each (promoter, gene, etc...)
		-where is the nucleosome positioned on each one (from the lit.)

	xls spread sheet of how the 16 nucleosome sequences compare
		-what is biologic function of each 
		-what did Pugh data have say about each sequence
		-where did the Pugh data locate the nucleosomes in each case


/autotmb/projects/bishop/TERAGRID
./Jiang-Pugh/Additional data file 1.xls
./JPdata/raw/Additional data file 1.xls
./JPdata/raw/Jiang-Pugh.xls
./JPdata/raw/Floe2010.Supp.xls
dna 113% pwd

There must be another file with all the array data giving the results from diff. experiments

graphical summary
	fuzziness,occupancy,biologic function, (type, gene) aand the measured positons


SOME USEFUL NOTES:
***********************
yeast genome link http://www.yeastgenome.org/ 
	you can make urls that go directly to your region of interest ifyou know the chromosome number and the sequence index 
	e.g.
	http://browse.yeastgenome.org/fgb2/gbrowse/scgenome/?name=chrI:158568..158714
		
Ian/Jason
	little fe bccd install worked I can now do MD on 3 or 4 of the nodes
	-you can do this too 
	I have some plans for this friday to work w/ some of Box's student to do the following:


1) get all nodes of little fe working
2) run some of the little fe demos to show that 1) has been achieved
3) install some missing software on little fe
     -xosview or some other graphical load monitoring software
    -missing libraries for VMD to run rlwrapper if I'm not mistaken
4) configure s.t. the head node of the little fe can run namd
   I have all the namd inputs it's just a matter of making some nice scripts to automate starting up a simulation
I have basic templates so real task is some tcsh, bash or other shell scripting.

5) An interesting alternative to running namd on head node is to plug my laptop into the switch and use my laptop as the visualization node.
the little fe head node is a dhcp server so I presume plugging in my laptop via ethernet gives it an IP address/config
and I can then ssh to any of hte other nodes on the little fe...
It seems this will be more of a user administration problem than anything else.


USEFUL LINKS FOR ALL OF THIS

namd:   www.ks.uiuc.edu
    or google: namd autoimd

little fe:  littlefe.net

bootable cluster cd: bccd.net


OTHER CLUSTER DISTROs: that I'd like to try/compare to bccd
***********************

NOTE: I've put various distros in /usr/local/Linux-Distros

but to make life simple I should probably provide another HD
or note that the scientific linux allow multiple distro installs.

rocks cluster: (alternative to bccd)
www.rocksclusters.org/

scientific linux: (alternative to bccd)
http://www.scientificlinux.org/