Yaamini’s Notebook: Demultiplexing Update

New MSConvert CLI, hot off the presses

Emma ran over a new version of msconvert Austin recoded to accomodate the idiosyncracies in our isolation windows. He generated a new .config file, found here. Below is a side-by-side comparison of two different .config files. The one on the left is the original code, and the one on the right is the modified code.


I used the following code to convert one .raw file:

 20170412_pwiz_testBuild_addMinWindowSize\msconvert.exe -c config_fix.txt 2017_January_23_envtstress_oyster1.raw  

If this works, I’ll start converting all of my files! Emma will also take my converted file and run it through PECAN using a modified isolation scheme. You can track my progress with this Jupyter Notebook.

One possible caveat is that Laura and I have different isolation schemes and isolation window sizes. It’s possible that the .config file I’m using won’t be compatible with Laura’s data, and is only specific to my data for this experiment.

from the responsible grad student http://ift.tt/2pc41pX

Hyak and PBJelly: Wall Time fixed

In the never ending excitement of learning how to use Hyak I learned something new about Slurm and sbatch.

sbatch is unable to take arguments via the command line, so if you try something like sbatch -s srlab -A srlab PBRun.sh --time 40:00:00 to run the PBRun.sh script from yesterday requesting 40 hours of run time, it will ignore the –time argument and run at the default of 1 hour. Not useful when you’re running a gap filling program that may take days to run.

A little bit of reading led me to realize that sbatch won’t accept arguments via the command line, instead requiring them to be passed via a shell/batch file that looks like

Screen Shot 2017-04-13 at 9.22.05 AM

After passing the batch file arguments to sbatch it seems to be running fine, with 10 days of max run time and 400gb of system ram requested on one node.

Screen Shot 2017-04-13 at 9.22.43 AM

So… progress? Now to see if we get an angry email from the Hyak admin staff for not optimally using our node.

Yaamini’s Notebook: Demultiplexing Data

Demultiplexing is perplexing

It’s time to modify our DIA Analysis pipeline and get more usable data from it. The first step would be to reconvert all of my .raw files to .mzmL files, this time demultiplexing the data.

When we set up m/z ratios for the mass spectrometer to use, the machine misses certain windows. To fix this, we add overlapping windows to our method. Demulitplexing allows us to separate overlapping windows.

Sean tried demultiplexing using the code on the DIA wiki, but ran into an error that referenced the isolation windows.

 msconvert.exe --zlib --64 --mzML --filter "peakPicking true 1-2" --filter "demultiplex optimization=overlap_only" *.raw  

I tried my hand at demultiplexing in this Jupyter notebook.

I used a .config file instead of directly inputting code because that’s what the Evernote suggested I do the last time I used MSConvert’s CLI.

The code I used, referencing my .config file:

 "c:\Program Files\ProteoWizards\ProteoWizard 3.0.10577\msconvert.exe" -c C:\Users\srlab\Documents\oystertest\msconvert-SIMMS1.config C:\Users\srlab\Documents\oystertest\2017_January_23_envtstress_oyster1.raw  

I got a different error from Sean. msconvert was reading my demultiplexing filter argument, but then ignoring it.


I tried several iterations of this, still getting errors, before I asked Emma what her lab suggested. The Github issue can be found here. Emma spoke with Austin, who is responsible for msconvert. Turns out, our isolation windows generated by Skyline have some quirks that prevent the program from recognizing them. Austin rewrote msconvert so it can read our files. I sent my files over to Emma to demultiplex since the new version is installed on the Genome Sciences server, so now I’m just waiting on the result!

from the responsible grad student http://ift.tt/2oAUV5c

Yaamini’s Notebook: Manchester Day 55

OA System Breakdown

Today Steven, Laura and I tied up loose ends from the OA exposure and prepared the system for the next month or so.

But first, some housekeeping from last week:

I didn’t look at the data Grace collected until today, but looks like we don’t have any conductivity information. This is not a huge deal because we think those values aren’t reliable anyways.


Figure 1. Water Chemistry Data from the last day of OA exposure

I also uploaded data from the oysters I sampled here. I plan on visualizing the data soon!

Alright, let’s get into what I did this week.

My first task was moving all of my oysters into two larger tanks so Laura could use the space she needed for her spawning set-up. I took the oysters in each tank and removed all epibionts and rinsed them with freshwater. I mainly removed barnacle and polychaete worms. I noticed that larger worms like to hide in the crevices between a larger oyster shell and a small attached shell. If I could safely remove a shell that was attached to a larger oyster and remove more polychaetes, I did. I put the epibionts in a bucket with bleachwater to sit until Laura returns to Manchester on Thursday.

I then split the oysters in each tank into two separate oyster bags.

Table 1. Number of oysters in each bag per tank.

Tank A B Total
1 8 9 17
2 9 8 14
3 7 7 14
4 9 8 17
5 9 9 18
6 8 8 16
Total 50 49 99

Bags labelled “A” went into the tank furthest from the dry lab, and bags labelled “B” went into the tank closest to the dry lab. We added a direct ambient line and airstone to each tank. All of these tanks were cleaned by Laura while I was removing epibionts.


Figure 2. Arrangement of holding tanks. The bottom left tank is Tank A, and the bottom right tank is Tank B. The top tank (closed, with lid) houses half of the oysters that were originally being held on the bottom shelf. The other half are in the yellow tray on the right, and were transported to SAFS.



Figures 3-4. Arrangement of oyster bags in Tanks A and B.

I then added a HOBO to each tank. HOBO 1 went into Laura’s conditioning tank, HOBO 2 into my original holding tank, HOBO 3 in Tank A, and HOBO 4 in Tank B.



Figures 5-6. HOBO loggers placed in Tank A and B.

Finally, I measured the flow rates for Tank A and B. They were approximately equal (Tank A: 1L in 1:23.34 min, Tank B: 1L in 1:25.98 min). The flow rate was roughly .75 L/min, which I figured would be enough. I will ask Laura to adjust flow on Thursday if the tanks look too gross.



Figures 7-8. Flow rates for Tanks A and B. They were both roughly .75 L/min.

Another productive day at Manchester! I don’t think I’ll be out there for a week or two until I have to do my heatshock. Ideally, I would like to learn how Laura is caring for her spawning individuals and larvae so I can do similar things when I spawn. For funsies, I’ll leave you with a picture of me with an oyster that looks like a heart.


from the responsible grad student http://ift.tt/2pal3BM