Mesostate

From CSBLwiki

Revision as of 07:22, 30 August 2010 (view source) Igchoi (Talk | contribs) (→Status(result)) ← Older edit		Revision as of 07:52, 30 August 2010 (view source) Igchoi (Talk | contribs) (→Status(result)) Newer edit →
Line 59:		Line 59:
	</pre>		</pre>
-	*Using [[R]] to check the data	+	*Using [[R]]
		+	<pre>
		+	meso = read.csv("tmp1.csv")
		+	table(meso$Mesostate) # space & underbar remove
		+	</pre>
		+
	<pre>		<pre>
	## load saved R data		## load saved R data
Line 89:		Line 94:
	*6 x 6 bins		*6 x 6 bins
	<pre>		<pre>
		+	library(ash)
		+	x = as.matrix(meso[,7:8])
		+	ab = matrix(c(-180,-180,180,180),2,2)
		+	nbin = c(6,6)
		+	bins = bin2(x,ab,nbin)
		+	-----
	> print(bins)		> print(bins)
	[,1]  [,2]    [,3]  [,4]  [,5]    [,6]		[,1]  [,2]    [,3]  [,4]  [,5]    [,6]

---

Revision as of 07:52, 30 August 2010

Contents 1 Concept 2 Procedure 2.1 Standard data set 2.2 Torsion angles 2.2.1 Mesostate 2.2.2 Calculation 2.2.3 Alphabet assignment 2.3 Profiling 2.4 Applications 3 Status(result) 4 References

Concept

Procedure

Standard data set

• We are going to use the SCOP DB: sequences and structures in the Astral compendium

Torsion angles

• Structural alphabets based on torsion angle distributions in the ramachandran plot

images to be posted...

Mesostate

• Data production by 배형섭
• Torsion angle mesostate by LINUS (Rose Lab)

Calculation

• Following tools can be used to calculate torsion angles of backbones
  • MMTK (python molecular modeling toolkit)
  • DSSP (by C. Sander)

Alphabet assignment

• Information theory

Profiling

• Normalization?
• Distance metric?

Applications

Status(result)

• Check your data (pdbstyle-1.75.new)
  • check.pl < pdbstyle-1.75.new > tmp.txt

while(<>) {
  chomp;
  @tmp = split/\t/,$_;
  if($tmp[8]=~/_/) { next }
  if(scalar(@tmp)==9) { print $_,"\n" }
}

• txt2csv.pl < tmp.txt > tmp1.csv

while(<>) {
  chomp;
  $_=~s/\[|\]//g;
  @line = split/\s+/,$_;
  $tmp = '"'.join("\"\,\"",@line)."\"";
  print $tmp,"\n";
}

• Using R

meso = read.csv("tmp1.csv")
table(meso$Mesostate) # space & underbar remove

## load saved R data
load("meso.rdata")
## analysis
dim(meso) # dimension 11,810,116 residues
meso[1:2,] # check first two rows in the data (list)
dom = unique(meso$Domain)
ndom = length(dom) # 65,485 SCOP domains
nrow(meso)/ndom # average 180 residues (domain size)
                # consider 1st, last residues are skipped..
##
## ramachandran plot
##
# randomly picking 5,000 residue's Phi & Psi
rn = sample(nrow(meso),5000)
png(file="ramachandran.plot")
plot(meso$Phi[rn],meso$Psi[rn],xlab="Phi",ylab="Psi",xlim=c(-180,180),ylim=c(-180,180),main="Ramachandran plot",col="gray")
# randomly picking 5,000 helices
rn = sample(which(meso$Structure=='H'),5000)
points(meso$Phi[rn],meso$Psi[rn],col="green")
# randomly picking 5,000 sheets
rn = sample(which(meso$Structure=='E'),5000)
points(meso$Phi[rn],meso$Psi[rn],col="red")
dev.off()
##

• 6 x 6 bins

library(ash)
x = as.matrix(meso[,7:8])
ab = matrix(c(-180,-180,180,180),2,2)
nbin = c(6,6)
bins = bin2(x,ab,nbin)
-----
> print(bins)
       [,1]  [,2]    [,3]   [,4]   [,5]    [,6]
[1,]  90545 18116   83067 120228 284531 1369188
[2,] 100305 84717 3685965 485872 653175 2066098
[3,]   4849 86683 1565006   5826  39358  294463
[4,]  23516  9348    2594 135671  25882    3456
[5,]  54491 12777  109935 234363  17244   41903
[6,]  35495  3524   13284   6641   5944   34984

References

1. Error fetching PMID 19188606: [FPP]