Laura’s Notebook: Day 50 Activities

Good news. We’ve actively used the OA system for 50 days with no major catastrophes. Celebrated by spending the day water sampling and cleaning oyster poop; here’s the summary:

  • Primed the pH probe with low pH water from tank 2
  • Labeled bottles
  • Picked up Rick’s Sond to measure salinity
  • Animals were being fed (header tank was ~3/4 full). Because I was planning on draining them in an hour or so, I turned off the flow from the header tanks to the culture tanks to let the oysters eat.
  • Took water samples @ 1pm
  • Measured grab samples for T, S & pH. I’ve been using both our salinity probe (even though it appeared broken, was trying to record uS/cm in case it was reading correctly), and Rick’s Sond, but today the glass bulb on our salinity probe broke off. I only recorded Salinity via the Sond. Finished measuring @ 1:45pm.
  • Injected HgCl into bottle samples; finished this @ 2:05pm.
  • Cleaned all tanks. Method:
    • Remove Olys and place in tray
    • Use large siphon to drain; when water level gets low I tip the tank on its edge to drain as much water as possible. Then I remove the probes if they are in there and place in a tri-pour with their respective treatment water.
    • Carry the tank to the floor, thoroughly spray gigas in tank with fresh water to remove large particles/feces. This time I counted the # of polychaets found in each tank, and collected the polychaets that I saw to give to Chelsea Wood – she is inspecting them to figure out which species it is. Most are found amongst the gigas – sitting on top of the valve that faces down, or in the cracks/crevaces. I found one in an Oly bag, and have seen a few in the past, but the majority are in gigas.
    • Dump gigas into tray.
    • Thoroughly spray each gigas and each oly back individually.
    • Sanitize the tank with Vortex cleaning solution, rinse with fresh water
    • Replace tank, replace cleaned animals.
    • I did not move the animals, tanks or probes around this time.
  • Bleached the algae lines with 20mL 12percent bleach in 3 gallons. Let bleach water run through system (from algae lines through the lines that run into the tanks – but redirected bleach water to the floor). Rinsed lines with 3 gallons of fresh water.
  • Replaced all filters


  • String/mucusy stuff re-appeared in tanks 1 & 2 in large amounts, and a small amount in tank 3. We need to sample this on Saturday if we see it.
  • No gigas mortality. I have not counted Oly mortality (there are definitely a few), but will do so on Saturday when we sample.

from The Shell Game

Marcos’ Trinity run: Completed! We…

Marcos’ Trinity run: Completed!

We finished Marcos’ Trinity assembly of his three flounder samples today, and have moved on to the TransDecoder step. Hoping to get everything done over the weekend so we can move on to Trinotate soon.

Jupyter Notebook found: here
3 Sample Assembly found: here
Trinity Assembly stats:


Hyak and You pt 3: Compilers.

As I mentioned before, Hyak comes with an older version of the GCC compiler, which isn’t compatible with Pitchfork, so I set out installing the new GCC Compiler. I forgot how long this takes, it’s been running 4+ hours now, and shows no sign of stopping. However, it *should* be installed in the shared srlab scratch directory, which would give everyone access to it, so we would only need to do it once. So… progress?

Fine flounder reproduction was induced…

Fine flounder reproduction was induced at the IMARPE laboratory. 10 individuals of the 3 (IMARPE-12), 40 (IMARPE-18) and 60 (IMARPE-19) days post-hatching were collected. Samples were stored in RNAlater. RNAzol was used for the isolation of the total RNA. DNA-free ™ DNA Removal Kit AM1906 was used to remove the DNA. RNA samples were sent in RNAstable columns for their sequencing. Three samples were sequenced in one lane.

RNA sequencing conditions were as follows:
Platform: HiSeq2500-High Throughput Mode
Library: Illumina TruSeq stranded mRNA
Read Length: Illumina 100bpPE
Number of reads: 100-150 millions / sample type

Results of RNA-Seq were as follows:

Screen Shot 2017-04-06 at 2.13.36 pm

Hyak and you pt 2. Github and Pitchfork.

We’re logged in and have a build node with internet access, so lets get some programs on here!

Pitchfork is a nice installer for PacBio related software, and since I plan to run PBJelly on Hyak, we’re going to need it.

First, we clone the GitHub repo via git clone pitchfork

which nets us an error stating:

Cloning into '/usr/lusers/seanb80/programs/pitchfork'...
Warning: Permanently added the RSA host key for IP address '' to the list of known hosts.
Permission denied (publickey).
fatal: Could not read from remote repository.

Please make sure you have the correct access rights
and the repository exists.

I know the repo exists, since I used it yesterday, so it’s an access thing. After a bit of googling, I realize that I forgot I had to add an SSH key for my GitHub account on the GitHub website. Oops.

ssh-keygen -t rsa -b 4096 -C ""

to make the key pairs

ssh-add -k ~/.ssh/id_rsa2

to add the code to the list of active key pairs

cat /usr/lusers/seanb80/.ssh/

to show the public key, which I copy in to GitHub via the directions found here starting at step 2.

We try the clone again and receive

Cloning into '/usr/lusers/seanb80/programs/pitchfork'...
remote: Counting objects: 9046, done.
remote: Total 9046 (delta 0), reused 0 (delta 0), pack-reused 9045
Receiving objects: 100% (9046/9046), 1.01 MiB | 0 bytes/s, done.
Resolving deltas: 100% (4769/4769), done.

[seanb80@n2185 pitchfork]$ ls
bin  LICENSE  Makefile	mk  ports

We installed something!

Then we switch to the legacy blasr version via git checkout legacy_blasr

and try make init

which gets us the error

[seanb80@n2185 pitchfork]$ make init
make -f mk/ init
make[1]: Entering directory `/gscratch/home/seanb80/programs/pitchfork'
/gscratch/home/seanb80/programs/pitchfork/bin/checkCC gcc
[ERROR] gcc version needs to be newer than or equal to 4.9
make[1]: *** [sanity] Error 1
make[1]: Leaving directory `/gscratch/home/seanb80/programs/pitchfork'
/gscratch/home/seanb80/programs/pitchfork/bin/checkCC gcc
[ERROR] gcc version needs to be newer than or equal to 4.9
make: *** [sanity] Error 1

What version of GCC are we running?

[seanb80@n2185 pitchfork]$ gcc --version
gcc (GCC) 4.8.5 20150623 (Red Hat 4.8.5-4)

Hm… Time to test out the yum package manager

[seanb80@n2185 pitchfork]$ yum install devtoolset-3-gcc-c++
Loaded plugins: fastestmirror
Repodata is over 2 weeks old. Install yum-cron? Or run: yum makecache fast
You need to be root to perform this command.

[seanb80@n2185 pitchfork]$ sudo yum install devtoolset-3-gcc-c++
bash: sudo: command not found

Hah. We don’t even have access to sudo on a build node. That’s an issue!

Found a workaround here

Hyak and you pt 1. Logging in.

I mentioned before there’s going to be a steep learning curve for Hyak, so I will attempt to document them both here, and on the forthcoming wiki page for LabDocs.

Hyak is generally like any other remote computer you can ssh into, with some caveats. First, there are different types of nodes you can be accessing at any point in time. There are login nodes (for logging in), build nodes(For building programs. These are the only nodes that have access to the internet), and execute nodes(the nodes for actual work).

Logging in using your UW NetID:

First, ask me either via email or in person to add you to the list of users for our node and then


in your favorite terminal application greets you with

Entrust passcode: 

The first password is your normal NetID password, the Entrust passcode is the number from the Entrust token you get by requesting it from here

After inputting both of those, you’re sitting on a login node. This has most of the normal commands you’re used to from Emu and Roadrunner, except it’s running Red Hat Enterprise as opposed to the Debian based Ubuntu you’re used to. This means no apt-get, instead we get to use yum.

Since we’ll be installing programs from GitHub and other places, we’ll need to switch to a build node

srun -p build --time=2:00:00 --pty /bin/bash

will request a build node be reserved for us for 2 hours. The --time argument for a build node maxes out at 8 hours, and some thought here regarding how long we’ll be actively using the build node is important, as these are shared resources and others may need them also.

Next on Hyak and You, Installing Pitchfork, Blasr, and PBJelly.

Hyak is live!

We got an email yesterday that we would have access to our Hyak node today at 9:00am. Exciting times, but there’s going to be a huge learning curve.

Here’s a teaser showing our node!

Screen Shot 2017-04-07 at 9.14.07 AM