[FEATURE] Add tandem_dup_count to Junction #147
Conversation
Irallia
requested review from
a team and
SGSSGene
and removed request for
a team
Codecov Report
@@ Coverage Diff @@
## master #147 +/- ##
==========================================
+ Coverage 94.75% 94.93% +0.17%
==========================================
Files 18 18
Lines 706 731 +25
==========================================
+ Hits 669 694 +25
Misses 37 37
Continue to review full report at Codecov.
|
src/structures/junction.cpp
Outdated
| @@ -27,11 +32,15 @@ bool operator<(Junction const & lhs, Junction const & rhs) | |||
| : lhs.get_mate2() != rhs.get_mate2() | |||
| ? lhs.get_mate2() < rhs.get_mate2() | |||
| : lhs.get_inserted_sequence() < rhs.get_inserted_sequence(); | |||
| // TODO (23.7.21, irallia): tandem_dup_count doesn't play a role here right? | |||
There was a problem hiding this comment.
@joshuak94 @joergi-w If the amount of duplications is bigger or smaller between junctions, does not play a role for size comparsion right? Just for equality.
There was a problem hiding this comment.
That would be weird...
In general, the following must hold: If a < b and b < a are both false, then a = b.
This would be violated if you omit the duplication count.
There was a problem hiding this comment.
Aaah, you mean a duplication is larger if it contains more duplications?
And thus, one junction is larger than another if this is the case.
Right, but we look before whether a junction is smaller from the position perspective (mate < mate). The question is, does it matter how high the tandem_dup_count is and if so where would you put that? Is that more relevant than the length of the inserted sequence?
There was a problem hiding this comment.
I guess, I haven't yet fully understood what a junction is...
Can two junctions with equal positions and sequence lengths possibly have a different duplication count? If yes, I would include it with the lowest priority.
My point is, if any of the fields in class junction differ, there must be a clear order (so you can only omit field comparisons that are indirectly determined, or that are omitted in the equality comparison as well).
There was a problem hiding this comment.
I have now thought about it again.
A junction describes a change in the read in comparison to the ref (e.g. insertion, duplication, deletion). It consists of two endpoints (breakends), a possibly inserted sequence and the name of the read and now also the number of duplications.
If we now compare two junctions, we want to know if they
- share approximately the same positions (breakend 1 and 2), then
- if they contain a similar number of duplications and
- the inserted sequence is of similar length.
Important to note, for a novel element insertion the inserted sequence is the whole new sequence, for a duplication it is the duplicated sequence, this can be shorter if there are multiple duplications. Therefore, I now compare first on duplication and then on the sequence.
The content of the sequence is not compared yet, but that is still planned.
Irallia
force-pushed
the
FEATURE/add_tandem_dup_count_to_junctions
branch
from
6867c64
to
48a7d05
Compare
include/structures/junction.hpp
Outdated
| @@ -13,6 +13,7 @@ class Junction | |||
| Breakend mate1{}; | |||
| Breakend mate2{}; | |||
| seqan3::dna5_vector inserted_sequence{}; | |||
| int16_t tandem_dup_count{}; | |||
There was a problem hiding this comment.
Is there a reason not to use uint16_t?
Intuitively a count is always positive.
test/api/clustering_test.cpp
Outdated
| @@ -14,6 +14,7 @@ int32_t const chrom1_position2 = 94734377; | |||
| int32_t const chrom1_position3 = 112323345; | |||
| std::string const chrom2 = "chr2"; | |||
| int32_t const chrom2_position1 = 234432; | |||
| int16_t tandem_dup_count = 0; | |||
There was a problem hiding this comment.
use const?
However, it would be good to write at least one test with more than 0 tandem duplications...
|
sorry, it has become a bit more with the test, because functionalities for clustering were also missing |
Signed-off-by: Lydia Buntrock <[email protected]>
…nt() function Signed-off-by: Lydia Buntrock <[email protected]>
Irallia
force-pushed
the
FEATURE/add_tandem_dup_count_to_junctions
branch
from
a125b1c
to
4df660e
Compare
Resolves first part of #143.
A characteristic of tandem duplications is the number of copies. I store this in the new variable
int16_t tandem_dup_count. It is 0 if there is no duplication (e.g. insertion (novel element), deletion, inversion ...).