severianovillarruel / deduper Goto Github PK

Hi Seve,
Awesome work!
Here are my comments for pseudocode
Does the proposed algorithm make sense to you? Can you follow the logic?
Yes, It is very clearly written and logic is easy to follow.

2. Does the algorithm do everything it's supposed to do? (see part 1)?
   Yes for the most part. In the pseudo-code for point # 7a , it is mentioned that "if the read is mapped to the forward strand" then it will check if the "tuple" is present in the forward dictionary, which makes sense.
   But, In point #8a, you are checking the same thing in the forward dictionary again, even if it is reverse mapped.I know you meant to check in the reverse dictionary and it is I think just a copy-paste error,but I would advise caution when writing actual code.
   Also, make sure you close your files!

3.Are proposed functions reasonable? Are they "standalone" pieces of code?
Yes, functions to parse cigar strings and flags are clearly written and standalone pieces of code.

Again, awesome work!
-Pranav

Hi Sevi.

I noted a way in which your pseudocode is an improvement over mine. I see that you include a step to not write-out a read if the bitwise flag indicates that it is not mapped. I'm going to add that to mine. :) Your pseudocode has high-level functions and appears quite logical. I think you might be able to get away with not having a reference dictionary outside of your loop. I'd be happy to show you how I plan on avoiding that, if you like. I am curious as to how you plan on doing the "flush/reset" of your global dictionary. Is that by nesting it inside a loop in such a way that additions get left behind, when you leave that loop? Also, I noted your example output and input sam files, but I did not see unit tests in your pseudocode. It's been really important to me to have unit tests for each function and for steps along the way, as I write my script. But, that may be because my scripting skills are still rudimentary.

Best, David

Seve,

This is a fantastic game plan! If anything, I was taking notes on what I can include in my own pseudocode! Your algorithm tracks the same as what I had in mind for this assignment, so it makes sense to me! However, I don't think you need two separate dictionaries to contain forward and reverse (I am am assuming it takes more computational capacity to do so). You could just save the tuple key to contain UMI + strand + chr + position and it would be specific enough to not warrant a new dictionary. Overall though, your included all the key points your algorithm was meant to accomplish! Nice Job! Your functions for parsing the flag and cigar string are great as well! I know I said I had a hard time understanding it, but I figured it out and it makes sense, I was just confused because you referred to "I" as "indexing" when I am pretty sure it indicates an "insert". Your pseudocode looks lovely! Keep up the good work! You are doing great!