Skip to content
This repository has been archived by the owner on Oct 28, 2022. It is now read-only.

Commit

Permalink
Making David's documentation changes
Browse files Browse the repository at this point in the history 
Also adding more pictures and hammering on the parent/child distinction.
  • Loading branch information
@adamnovak
adamnovak committed Feb 6, 2015
1 parent 5e43248 commit 5efdc7b
Showing 1 changed file with 67 additions and 34 deletions.
101 changes: 67 additions & 34 deletions src/main/resources/avro/common.avdl
View file
Original file line number Diff line number Diff line change
Expand Up @@ -26,23 +26,23 @@ nicks in it, one on the top strand and one on the bottom strand. Here is an
example where the top strand is GGTGGNG.
```
| | <- top strand nick at position (2,+)
5' ------- ------------------ 3'
| <- top strand nick at position (2,+)
5' ------- ------------------- 3'
G G T G G N G
C C A C C N C
3' --------------- ---------- 5'
position (3,-) -> | | in bottom strand nick
3' ---------------- ---------- 5'
| <- bottom strand nick at position (3,-)
+0- +1- +2- +3- +4- +5- +6- <- coordinates
```
A sequence is a piece of double stranded DNA composed of a series of DNA
A sequence is a piece of double-stranded DNA composed of a series of DNA
basepairs. In the default forward orientation a sequence is specified by the DNA
letters of its top strand, e.g. in this case GGTGGNG, where N indicates an
unknown base. The basepairs in this example sequence are indexed left to right
from 0 to 6 relative to this default orientation. A sequence containing regions
of uncertainty, denoted by Ns, is called a scaffold, while a sequence in which
every base is known is called a contig. This sequence is a scaffold.
unknown base. The basepairs are indexed left to right from 0 to 6 relative to
this default orientation. A sequence containing regions of uncertainty, denoted
by Ns, is called a scaffold, while a sequence in which every base is known is
called a contig. This sequence is a scaffold.
In its default forward orientation, each basepair has a left or "+" side and a
right or "-" side. For example, the left side of the T/A basepair in the above
Expand All @@ -54,15 +54,15 @@ example, the right side of the following G/C base pair is represented by the
position (3,-).
One way to think about a position in a sequence is to imagine that the DNA
double helix is nicked, e.g. as shown in the top strand DNA nick above. When
double stranded DNA is nicked, the 5'-3' phosphodiester bond between two
adjacent bases on one strand is broken, leaving an exposed hydroxyl group
(denoted OH) on the 3' side and an exposed phosphate group on the 5' side
(denoted -PO4). From a chemical point of view, you can think of a position in a
sequence as the the location in the double-stranded DNA of the exposed 5'
phosphate group of a nick. The position (2,+) is the -PO4 part of the top strand
nick in the diagram above, and the position (3,-) is the -PO4 part of the bottom
strand nick.
double helix is nicked, as shown above. When double stranded DNA is nicked, the
5'-3' phosphodiester bond between two adjacent bases on one strand is broken,
leaving an exposed phosphate group on the 5' side, shown by the vertical bar
(and an exposed hydroxyl group on the 3' side, not shown). From a chemical point
of view, you can think of a position in a sequence as the the location in the
double-stranded DNA of the exposed 5' phosphate group created by a nick. The
position (2,+) is the exposed phosphate group of the top strand nick in the
diagram above, and the position (3,-) is the same thing in the bottom strand
nick.
Each nonempty sequence has a "start", which is the left side of the first
basepair, and an "end", which is the right side of the last basepair. Each
Expand Down Expand Up @@ -104,12 +104,12 @@ configuration such that the middle piece is inverted and a new C/G basepair is
inserted at the first break, so the result would look like this:
```
| | | | | |
5' ------- - ----- ---------- 3'
| | |
5' ------- -- ------ ----------- 3'
G G C C A G N G
C C G G T C N C
3' ------- - ----- ---------- 5'
| | | | | |
3' -------- -- ------ ---------- 5'
| | |
+0- +1- ? -3+ -2+ +4- +5- +6- <- coordinates
```
Expand All @@ -123,7 +123,7 @@ be defined to carry it. Let us call the existing GGTGGNG reference sequence
```
"ref": "nov1": "ref":
- 3' 5' - 3' 5' -
... G C C ...
... C G G ...
Expand All @@ -149,16 +149,16 @@ endJoin = (3, -), "ref"
```
To describe the other new phosphodiester bond that needs to be added to the
To describe the other new phosphodiester bonds that need to be added to the
graph, between (2,+) and (4,+) in "ref", we need to create a new reference
sequence which contains no basepairs at all; it is the empty sequence of DNA
basepairs. Let the reference sequence "nov2" be such a reference sequence, with
start and end joins set to describe this phosphodiester bond. The reference
sequence "nov2" has no coordinates, and no positions. It looks like this
start and end joins set to describe these bonds. The reference sequence "nov2"
has no coordinates, and no positions. It looks like this
```
"ref": "nov2": "ref":
- 3' 5' 3' 5' -
... A G ...
... T C ...
Expand All @@ -167,7 +167,7 @@ sequence "nov2" has no coordinates, and no positions. It looks like this
-2+ +4- <- coordinates
```
An reference sequence with only start and end joins is describable by a segment
A reference sequence with only start and end joins is describable by a segment
with only start and end joins. A segment from the "nov2" reference sequence
would look like this:
Expand Down Expand Up @@ -218,6 +218,16 @@ end = null
endJoin = (4,+), "ref"
```
It looks like this:
```
--C--
| |
=G==G==T==G==G==N==G=
| |
-----
```
In this graph, there are two paths from (0,+) on "ref" to (6,-) on the same
sequence. One of them proceeds directly along "ref", and represents the un-
rearranged condition, while one detours through "nov1" and "nov2", and
Expand All @@ -236,6 +246,16 @@ end = (1,-)
endJoin = null
```
It looks like this:
```
--C--
| |
=G>=G=>T=>G=>G>=N>=G>
| |
-----
```
The rearranged path is describable by a list of 5 segments, 3 of which are the
contiguous pieces of the "ref" sequence, and 2 of which describe the newly added
"nov1" and "nov2" sequences:
Expand All @@ -248,7 +268,8 @@ length = 2
end = (1,-)
endJoin = null
2. Segment with top strand C in the "nov1" reference sequence:
2. Segment with top strand C in the "nov1" reference sequence, joined to parent
"ref":
startJoin = (1,-), "ref"
start = (0,+)
length = 1
Expand All @@ -262,7 +283,8 @@ length = 2
end = (2,+)
endJoin = null
4. Segment with top strand empty in the "nov2" reference sequence:
4. Segment with top strand empty in the "nov2" reference sequence, joined to
parent "ref":
startJoin = (2,+), "ref"
start = null
length = 0
Expand All @@ -277,17 +299,28 @@ end = (6,-)
endJoin = null
```
It looks like this:
```
--C>-
^ |
=G>=G=<T=<G==G>=N>=G>
| ^
-->--
```
Notice that we always maintain the coordinate system on the reference sequence.
The inverted middle segment is specified as starting at a position corresponding
to a nick in the bottom strand of the reference sequence, and continuing on 2
basepairs in a right-to-left direction along the reference sequence. It does not
get new coordinates relative to its final orientation or position in the
rearranged conformation. Also, all of the novel adjacencies in the sequence
graph involve actual novel segments, even if some of these are empty. This is
good for bookkeeping. Overall, this scheme allows us to create sequence graphs
describing any number of configurations and reconfigurations of reference DNA
sequences, and to describe paths through those graphs representing particular
arrangements.
good for bookkeeping; we want to be able to keep the parent sequences constant
and add new child sequences to express new variants. Overall, this scheme allows
us to create sequence graphs describing any number of configurations and
reconfigurations of reference DNA sequences, and to describe paths through those
graphs representing particular arrangements.
*/
protocol Common {

Expand Down

0 comments on commit 5efdc7b

Please sign in to comment.