[S] Getting it together

[sr320]

• [x] Read Git for Scientists 67-83
• [x] What are your opinions of Git and do you think this reading provides valuable insight / tips regarding your workflow?
• [x] How is the computational work going this past week, what are the highlights and biggest challenges faced?

[SMay1]

I was initially skeptical of Git because I felt that the barrier to entry was high. I thought it would take a lot of work to learn and wasn't sure of the benefits... But now I think everyone should be using Git. Version control is the kind of thing you hope you never need to use, but will be very glad you have when you need it.

I don't think that I will ever work on coding project that requires multiple collaborators to all be committing and merging edits and things- but for the more simple uses of Git, I think I will try and use it as much as possible from now on.

The reading was mostly review, although honestly I haven't been using Git at the command line much. I am only working in a single project directory and GitHub Desktop has been working well for my purposes- although it is nice to know the option is there for more complicated things. It was also nice to understand the workflow behind Git, as I wasn't really aware of "staged" files vs "tracked" files.

A couple questions:

• How often does everyone actually commit? Do you commit multiple times as you are working on a single file? I have only been committing maybe once or twice a day as I work on things
• Should I be cloning directories that I am working on frequently? (aka the GT-Seq GitHub page). I have really just been copying scripts from there instead of cloning them via github... is there any real benefit of cloning someone else's directory?

[SMay1]

Computational work for my project has been going great- in fact, I finished the first round of optimization for my GT-Seq panel just yesterday (hence this last minute entry). There have been a lot of steps involved, and organizing my workflow/ directories has been extremely helpful to making the whole process more efficient.

The biggest challenge was honestly knowing when to stop... I have been trying all kinds of different angles to identify problematic loci in my panel, and at some point I just needed to say "it is good enough." I am very happy with how it looks now.

Essentially I am done with Bioinformatics for the time being until I get more sequencing data back two weeks from today.

One thing that I want to try to do is get away from using Evernote and try to exclusively use Jupyter Notebooks. I think that for the next optimization round (which will go much faster because I know what I am doing and have the work flow ready to go) I would like to try to just use Jupyter Notebooks to keep all of my notes/ figures/ scripts etc... I might not like it, but I am going to give it a go.

[SMay1]

Network Plot of Cross-Amplifiers

Example of a 'good' locus:

Example of some 'bad' loci:

I ended up tossing out 30 loci of the 233 originally in my panel- leaving me with 203 loci. I have one more round of optimization to do, which will hopefully look great- and I will be ready to sequence 4,000 individuals!