Progress bae

From CSBLwiki

Revision as of 01:22, 21 March 2011 (view source) Gnusnah (Talk | contribs) ← Older edit		Latest revision as of 05:17, 12 August 2012 (view source) Gnusnah (Talk | contribs) (→2011)
Line 4:		Line 4:
	==2011==		==2011==
-	*Reference: [[media:JMB_adda.pdf|PDF download]]	+	*Reference:
	<biblio>		<biblio>
	ADDA pmid=12706730		ADDA pmid=12706730

---

Latest revision as of 05:17, 12 August 2012

Contents 1 2011 2 Schedule 3 ADDA algorithm 3.1 How to get residue correlation matrix

2011

• Reference:

1. Error fetching PMID 12706730: [ADDA]

Schedule

ADDA algorithm

How to get residue correlation matrix

• sample data: sample.tab

# set working directory
setwd("/Users/igchoi/Downloads/")

# read blast table
tmp = read.table("sample.tab",sep="\t")

# check dimension of table
dim(tmp)

# change simpler GI number (skip this if you don't understand)
tmp[,1] = sapply(as.vector(tmp[,1]), function(x) {
  y = strsplit(x,"\\|"); return(unlist(y)[2]) })
tmp[,2] = sapply(as.vector(tmp[,2]), function(x) {
  y = strsplit(x,"\\|"); return(unlist(y)[2]) })

# check the result
tmp[1:5,]

# aligned region: V7 V8, V9 V10
# query length = 590 residues
count.que = matrix(0,nrow(tmp),590)  # <- build empty 'aligned' matrix
for(i in 1:nrow(tmp)) {
  res = unlist(tmp[i,7:8])
  print(res)
  count.que[i,c(res[1]:res[2])] = 1
}

# check aligned region (blue colored regions are aligned with other proteins - central region)
image(t(count.que),col=c("white","blue"))

# count number of neighbors occurring both position i and j (using 'aligned' matrix)
corr = matrix(0,590,590)  # <- set empty correlation matrix (590 x 590)

for (i in 1:ncol(count.que)) {
  for (j in i:ncol(count.que)) {
    print(paste(i,j))
    chk = count.que[,i]+count.que[,j]
    cnt = length(which(chk==2))
    corr[i,j] = cnt
  }
}

# range of number of neighbors (among 52 hits)
range(corr)

# check highly correlated region in the correlation matrix
image(corr,col=topo.colors(10))