That Ex-Redditor who somehow speedran Tumblr (Posts tagged at university)

1.5M ratings
277k ratings

See, that’s what the app is perfect for.

Sounds perfect Wahhhh, I don’t wanna
felixcloud6288
felixcloud6288

So I discovered I had accidentally deleted the directory to a research project and only realized after every backup of it was gone.

I've been scrambling for the last few hours to redownload all the data files and tools, salvage whatever reports I had already sent to my professor, and rebuild as much of the codebase of the most immediately important things from what I can remember.

At this point, I can feel my heart about to explode and I feel like I'm about to puke.

felixcloud6288

As of this moment, I've managed to recollect all the files and tools I used in the research and have gotten most of the processed data I've been analyzing.

While rebuilding the processing pipeline, I discovered a few things about how I ran it initially that I'm going to bring up to my professor because maybe I can get some better or more diverse results using what I found.

felixcloud6288

I’m now getting way better results than I got previously. Either I did something horribly wrong when rebuilding the codebase, or there was a catastrophic bug in the old version that I failed to replicate in the new one.

original post at university
felixcloud6288
felixcloud6288

So I discovered I had accidentally deleted the directory to a research project and only realized after every backup of it was gone.

I've been scrambling for the last few hours to redownload all the data files and tools, salvage whatever reports I had already sent to my professor, and rebuild as much of the codebase of the most immediately important things from what I can remember.

At this point, I can feel my heart about to explode and I feel like I'm about to puke.

felixcloud6288

As of this moment, I’ve managed to recollect all the files and tools I used in the research and have gotten most of the processed data I’ve been analyzing.

While rebuilding the processing pipeline, I discovered a few things about how I ran it initially that I’m going to bring up to my professor because maybe I can get some better or more diverse results using what I found.

original post at university

So I discovered I had accidentally deleted the directory to a research project and only realized after every backup of it was gone.

I’ve been scrambling for the last few hours to redownload all the data files and tools, salvage whatever reports I had already sent to my professor, and rebuild as much of the codebase of the most immediately important things from what I can remember.

At this point, I can feel my heart about to explode and I feel like I’m about to puke.

original post at university
felixcloud6288
felixcloud6288

What was the cause of my coding issue?

A bug in my code

An undefined behavior when writing to multiple files in parallel

A memory leak

Overloading the system's memory

I was looking in the wrong output file

See Results
felixcloud6288

Situation: I’m writing multiple output files tracking the overlap of several thousands of genetic reads and organizing them by cell type. When doing a few trial runs, the files had tons of blank lines and the overlap scores weren’t rounding properly. I checked multiple times why nothing I was doing was fixing the issue. I also had to implement a check where if my computer reached >80% memory use, the program would run an emergency write of all the results it currently had.

Cause of the issue:

Keep reading

original post at university

Been spending weeks trying to learn and work with a tool called cellranger to understand how to process fastq files into count matrices of chromatin access regions and I’ve just hit an impassible snag.

Cellranger is a processing tool that takes fastq files produced by an Illumina sequencer and it can only work if provided all the output files from the sequencer, including files about the primers which aren’t available in almost every publicly available dataset because they have no actual value for research purposes.

I was able to rebuild one of those primer files based on the headers of each read in the files I’ve been working with, but it’s impossible for me to figure out what the other file’s contents would have been.


So anyway, I’m going to have to start this thing all over. Hopefully I can find a similar tool that doesn’t have such frustrating specifications.

original post at university

Learning a new Bioinformatics program is like learning to ride a bicycle if the bicycle came disassembled and it didn’t come with half the parts you needed and you have to use bolts that require a very specific type of wrench you can only get if you go to a specific store that demands you smelt the steel they need to make it and then the bicycle turns out to be useless because the wheels are square.

original post at university