Figure 3. Parallel versions of Glimmer Process workflow
Figure 4. Parallel versions of TigrScan process work flow
Pseudo Code for Database segmentation
#Open and read the genome file
Read $genomefile
#Read the 1st line of the FASTA file and display
Read line 1 of $genomefile
Display line1
#Extract the number of bases per line
$bases : no. of bases per line
$zone <-- f(organism,$bases)
Input : $THREADS
#Read the number of lines
$line : no. of lines in the $genomefile
#Compute the distribution desired
$split_lines <-- $lines/$THREADS
$firstline <-- 1
#Remove the existing splitfiles and intersection files
remove all splitfiles
remove all intersection files
#Main splitting code
for $i = 1 to $THREADS-1
{
open splitfile$I
if $i=1 then ($previous <-- $THREADS) else ($previous <-- $i-1)
open intersection$i
open intersection$previous
$lowerzone <-- ($i*$split_lines)-$zone
$upperzone <-- (($i-1)*$split_lines)+$zone
for $line = $firstline to $firstline+$split_lines
{
write to splitfile$i
if ($line > $lowerzone) write to intersection$i file
if ($line <= $upperzone) write to intersection$previous file
}
$firstline <-- $line
#This section ensures that no line gets missed in the
#last thread file , since the average distribution estimated
# per line ($split_lines) may not be an integer
open splitfile$THREADS
$previous <-- $THREADS-1
open intersection$THREADS file
open intersection$previous file
$lowerzone <-- $lines - $zone
$upperzone <-- (($THREADS -1)* $split_lines) + $zone
while (presence of line from current position in $genomefile)
{
$line <-- $line + 1
write to splitfile$THREADS
if $line>$lowerzone write to intersection$THREADS file
if $line<=$upperzone write to intersection$previous file
}
#This section builds the last intersection
#file in order as a FASTA format
$line <-- 0
open intersection$THREADS file
open dummy file
while(presence of line in intersection$THREADS)
{
$line <-- $line + 1
if $line>$zone write to dummy
}
$line <-- 0
while(presence of line in intersection$THREADS)
{
$line <-- $line + 1
if $line<=$zone write to dummy
}
remove intersection$THREADS
rename dummy as intersection$THREADS
exit