<s>
Phrap	B-Application
is	O
a	O
widely	O
used	O
program	O
for	O
DNA	O
sequence	O
assembly	O
.	O
</s>
<s>
It	O
is	O
part	O
of	O
the	O
Phred-Phrap-Consed	O
package	O
.	O
</s>
<s>
Phrap	B-Application
was	O
originally	O
developed	O
by	O
Prof	O
.	O
Phil	O
Green	O
for	O
the	O
assembly	O
of	O
cosmids	O
in	O
large-scale	O
cosmid	O
shotgun	O
sequencing	O
within	O
the	O
Human	O
Genome	O
Project	O
.	O
</s>
<s>
Phrap	B-Application
has	O
been	O
widely	O
used	O
for	O
many	O
different	O
sequence	O
assembly	O
projects	O
,	O
including	O
bacterial	O
genome	O
assemblies	O
and	O
EST	O
assemblies	O
.	O
</s>
<s>
Phrap	B-Application
was	O
written	O
as	O
a	O
command	O
line	O
program	O
for	O
easy	O
integration	O
into	O
automated	O
data	O
workflows	O
in	O
genome	O
sequencing	O
centers	O
.	O
</s>
<s>
For	O
users	O
who	O
want	O
to	O
use	O
Phrap	B-Application
from	O
a	O
graphical	O
interface	O
,	O
the	O
commercial	O
programs	O
MacVector	B-Application
(	O
for	O
Mac	B-Operating_System
OS	I-Operating_System
X	I-Operating_System
only	O
)	O
and	O
CodonCode	B-General_Concept
Aligner	I-General_Concept
(	O
for	O
Mac	B-Operating_System
OS	I-Operating_System
X	I-Operating_System
and	O
Microsoft	B-Application
Windows	I-Application
)	O
are	O
available	O
.	O
</s>
<s>
A	O
detailed	O
(	O
albeit	O
partially	O
outdated	O
)	O
description	O
of	O
the	O
Phrap	B-Application
algorithms	O
can	O
be	O
found	O
in	O
the	O
.	O
</s>
<s>
A	O
recurring	O
thread	O
within	O
the	O
Phrap	B-Application
algorithms	O
is	O
the	O
use	O
of	O
Phred	O
quality	O
scores	O
.	O
</s>
<s>
Phrap	B-Application
used	O
quality	O
scores	O
to	O
mitigate	O
a	O
problem	O
that	O
other	O
assembly	O
programs	O
had	O
struggled	O
with	O
at	O
the	O
beginning	O
of	O
the	O
Human	O
Genome	O
Project	O
:	O
correctly	O
assembling	O
frequent	O
imperfect	O
repeats	O
,	O
in	O
particular	O
Alu	O
sequences	O
.	O
</s>
<s>
Phrap	B-Application
uses	O
quality	O
scores	O
to	O
tell	O
if	O
any	O
observed	O
differences	O
in	O
repeated	O
regions	O
are	O
likely	O
to	O
be	O
due	O
to	O
random	O
ambiguities	O
in	O
the	O
sequencing	O
process	O
,	O
or	O
more	O
likely	O
to	O
be	O
due	O
to	O
the	O
sequences	O
being	O
from	O
different	O
copies	O
of	O
the	O
Alu	O
repeat	O
.	O
</s>
<s>
Typically	O
,	O
Phrap	B-Application
had	O
no	O
problems	O
differentiating	O
between	O
the	O
different	O
Alu	O
copies	O
in	O
a	O
cosmid	O
,	O
and	O
to	O
correctly	O
assemble	O
the	O
cosmids	O
(	O
or	O
,	O
later	O
,	O
BACs	O
)	O
.	O
</s>
<s>
However	O
,	O
Phrap	B-Application
does	O
not	O
rule	O
out	O
such	O
alignments	O
entirely	O
,	O
and	O
the	O
cross_match	O
alignment	O
gap	O
and	O
alignment	O
penalties	O
used	O
while	O
looking	O
for	O
local	O
alignments	O
are	O
not	O
always	O
optimal	O
for	O
typical	O
sequencing	O
errors	O
and	O
a	O
search	O
for	O
overlapping	O
(	O
contiguous	O
)	O
sequences	O
.	O
</s>
<s>
Phrap	B-Application
attempts	O
to	O
classify	O
chimeras	O
,	O
vector	O
sequences	O
and	O
low	O
quality	O
end	O
regions	O
all	O
in	O
a	O
single	O
alignment	O
and	O
will	O
sometimes	O
make	O
mistakes	O
.	O
</s>
<s>
Furthermore	O
,	O
Phrap	B-Application
has	O
more	O
than	O
one	O
round	O
of	O
assembly	O
building	O
internally	O
and	O
later	O
rounds	O
are	O
less	O
stringent	O
-	O
Greedy	O
algorithm	O
.	O
</s>
<s>
Phrap	B-Application
appears	O
error	O
prone	O
in	O
comparison	O
with	O
newer	O
assemblers	O
like	O
Euler	O
and	O
cannot	O
use	O
mate-pair	O
information	O
directly	O
to	O
guide	O
assembly	O
and	O
assemble	O
past	O
perfect	O
repeats	O
.	O
</s>
<s>
Phrap	B-Application
is	O
not	O
free	O
software	O
so	O
it	O
has	O
not	O
been	O
extended	O
and	O
enhanced	O
like	O
less	O
restricted	O
open-source	O
software	O
Sequence	O
assembly	O
.	O
</s>
<s>
Another	O
use	O
of	O
Phred	O
quality	O
scores	O
by	O
Phrap	B-Application
that	O
contributed	O
to	O
the	O
program	O
's	O
success	O
was	O
the	O
determination	O
of	O
consensus	O
sequences	O
using	O
sequence	O
qualities	O
.	O
</s>
<s>
In	O
effect	O
,	O
Phrap	B-Application
automated	O
a	O
step	O
that	O
was	O
a	O
major	O
bottleneck	O
in	O
the	O
early	O
phases	O
of	O
the	O
Human	O
Genome	O
Project	O
:	O
to	O
determine	O
the	O
correct	O
consensus	O
sequence	O
at	O
all	O
positions	O
where	O
the	O
assembled	O
sequences	O
had	O
discrepant	O
bases	O
.	O
</s>
<s>
This	O
approach	O
had	O
been	O
suggested	O
by	O
Bonfield	O
and	O
Staden	O
in	O
1995	O
,	O
and	O
was	O
implemented	O
and	O
further	O
optimized	O
in	O
Phrap	B-Application
.	O
</s>
<s>
Basically	O
,	O
at	O
any	O
consensus	O
position	O
with	O
discrepant	O
bases	O
,	O
Phrap	B-Application
examines	O
the	O
quality	O
scores	O
of	O
the	O
aligned	O
sequences	O
to	O
find	O
the	O
highest	O
quality	O
sequence	O
.	O
</s>
<s>
In	O
the	O
process	O
,	O
Phrap	B-Application
takes	O
confirmation	O
of	O
local	O
sequence	O
by	O
other	O
reads	O
into	O
account	O
,	O
after	O
considering	O
direction	O
and	O
sequencing	O
chemistry	O
.	O
</s>
<s>
If	O
a	O
consensus	O
base	O
is	O
covered	O
by	O
both	O
high-quality	O
sequence	O
and	O
(	O
discrepant	O
)	O
low-quality	O
sequence	O
,	O
Phrap	B-Application
's	O
selection	O
of	O
the	O
higher	O
quality	O
sequence	O
will	O
in	O
most	O
cases	O
be	O
correct	O
.	O
</s>
<s>
Phrap	B-Application
then	O
assigns	O
the	O
confirmed	O
base	O
quality	O
to	O
the	O
consensus	O
sequence	O
base	O
.	O
</s>
<s>
Phred	O
and	O
Phrap	B-Application
,	O
and	O
similar	O
programs	O
who	O
picked	O
up	O
on	O
the	O
ideas	O
pioneered	O
by	O
these	O
two	O
programs	O
,	O
enabled	O
the	O
assembly	O
of	O
large	O
parts	O
of	O
the	O
human	O
genome	O
(	O
and	O
many	O
other	O
genomes	O
)	O
at	O
an	O
accuracy	O
that	O
was	O
substantially	O
higher	O
(	O
less	O
than	O
1	O
error	O
in	O
10,000	O
bases	O
)	O
than	O
the	O
typical	O
accuracy	O
of	O
carefully	O
hand-edited	O
sequences	O
that	O
had	O
been	O
submitted	O
to	O
the	O
GenBank	O
database	O
before	O
.	O
</s>
