Peer Review - Chris Chua #1
Description
Does the proposed algorithm make sense to you? Can you follow the logic?
The algorithm makes sense. There are some serious, but easily fixable flaws in the logic.
Please see questions below.
Does the algorithm do everything it's supposed to do? (see part 1)
The algorithm does everything it is supposed to do in Part 1.
I appreciate that the code keeps all removed reads in a discarded output file.
Are proposed functions reasonable? Are they "standalone" pieces of code?
The functions are reasonable. I would combine position_fw and _rv.
See part 3 to see if any of main code would benefit from being made into functions if additional functionality is added.
###SPECIFICS###
position_fw:
- else statement is returning the wrong variable
- how are you pulling the number of bases to adjust the position from the CIGAR string?
position_rv:
- would the soft-clipping on the right side of the sequence affect the left-most position (even if it is read 2)?
main code:
- would you also write out the @ header lines to the other output files?
- how are you storing the columns after being separated?
- how are you retaining position information as you check between lines? saving values to variables?
- are reads with the same position but different strandedness duplicates?
- It appears that reads are only considered duplicates by the rname, I would also require the UMI, position, and strandedness to be the same.
(this is because I do not know if samtools also sorts the reads by UMIs or just by leftmost position)
Suggestions:
- wrap text with newline characters so you don't have to scroll left
Let me know if you have any questions or concerns! Also if you would like me to review your pseudocode again.