Grace’s Notebook: My Bioanalyzer and Sam’s RNA Clean-up Kit and RNeasy Plus Mini Kit results

Today I ran the Bioanalyzer on the samples Sam isolated RNA from months ago using RNAzol RT as well as some of the RNA isolated from samples that I have done. The results were not good – no dye showed up, so maybe I did it wrong. TBD. Also, Sam used a kit to clean up the RNA that I have isolated and then ran Qubit and Nanodrop1000, as well as used RNeasy Plus Mini Kit to isolate RNA from untouched hemolymph samples. Low concentrations of RNA, but at least there is RNA. The results from the RNeasy Plus Mini Kit isolations are clean and the RNA on Qubit and Nanodrop1000 are good.

Bioanalyzer (GitHub Issue #328)

Used RNA Pico Chip. I ran the three samples that Sam isolated RNA from in November 2017. I also ran the following 8 samples from what I previously isolated RNA from:

I loaded the RNA pico chip according to the protocol. Vortex the chip (first time through, it didn’t click in to the tension bar very well and popped out. I ran it anyway, but this error popped up:)

Use 2100 Expert software on the desktop.

Assay selection: RNA –> Eukaryote Total RNA Pico series II
Save –> Custom –> Desktop\data\RobertsLab
After the chip is loaded, name the samples






Results aren’t great. Sam noted that:
A couple of things: No marker is showing up in any of the wells. Expect a tall, narrow, defined peak at ~20 - 23 seconds. These have not been DNased, so it may not be surprising that these look crazy?

Sam Cleans up Tanner Crab RNA (GitHub Issue #330)

Link to his notebook post: here

Essential results:

  • All concentrations were too low for detection via NanoDrop.
  • Qubit quantification indicate yields ranging from ~25ng to ~192.5ng.
  • New Qubit results are much lower (except for tubes 409 and 355)

    Here’s how his new cleaned results compare to my previous Qubit results:

    My previous results (column highlighted in pink compares to the blue highlighted column in Sam’s Qubit results:)

Sam’s results

Sam isolates RNA from hemolymph samples using RNeasy Plus Mini Kit (GitHub Issue #327)

Link to his notebook post: here

I gave Sam four samples from Day 26 (samples were taken in triplicates) from the following crabs:
451 (infected, cold, immature)
478 (infected, ambient, immature)
506 (uninfected, ambient, immature)
432 (uninfected, cold, mature)

Essential results:

  • NanoDrop results look good (purity is good)
  • Qubit yields are decent (range from ~37ng to ~350ng)

Sam notes that the important thing is that this procedure produced clean RNA, which renders the Qubit results believable. He thinks the RNAzol RT isolations I did had too much contamination carried over causing incorrect Qubit measurements.

I’m not sure if that means I wasn’t careful enough with not carrying contamination between samples or if RNAzol RT didn’t work well with the salts and other junk in the crab hemolymph…

Regardless, we’ll try using TriReagent out as soon as some arrives and ponder the idea of getting some kind of isolation kit (Sam’s suggestion).

from Grace’s Lab Notebook

Laura’s Notebook: August 2018 goals

  1. Advise Emily & Micah on how to sample deployed Olys
  2. Submit geoduck paper
  3. Submit Polydora paper
  4. Write O. angasi conditioning experiment methods
  5. Embed, section and mount O. angasi histology tissues
  6. Plan my attack for bringing samples back to USA: Mail or checked luggage?
  7. PCSGA talk
  8. Headway on 2018 Oly paper results (Kaitlyn doing lots of leg work)
  9. Headway on 2017 Oly paper results (Decisions need to be made, incorporate eelgrass deployment data?)

from The Shell Game

Grace’s Notebook: Trying out RNA protocol with 6x original protocol’s volumes of reagents

Today Steven, Sam, and I met to talk about the issue with poor results on the QUbit and Nanodrop1000 on the pooled hemolymph samples I prepared. Short-term plan is to try out the RNA protocol, but multiply all reagents times 6 (so, use 6ml of RNAzol, etc.) to see if that ratio would help; run the Bioanalyzer on the samples that Sam processed as well as some random samples that I isolated. Today, I did the 6x RNA protocol on 4 of the extra samples from day 26 (samples from day 26 were taken in triplicate).

RNA Isolation Protocol x 6

Link to the protocol I did today: HERE
*Note: I forgot to do step 6

It was a bit challenging today because in order to support those larger volumes of reagents, I had to use 13ml round-bottomed tubes. It was difficult to see any pellets or genetic material, but I did my best. I oriented the tubes in the same way every time I put them in the centrifuge and I marked the orienation on the tubes so that I knew where the genetic material should be if it was going well.

Per Sam’s suggestion, I resuspended the samples in 25ul of 01% DEPC-treated H20, instead of the typical 50ul. I also pipetted to mix/dissolve.

I ran the Qubit, and they were ALL “out of range/too low”.

I did not run the Nanodrop, because I can’t remember how to use it. I will do it first thing tomorrow when Sam is in.

Goals for tomorrow:

  • Run Nanodrop on the samples I isolated yesterday
  • Run Bioanalyzer on the samples Sam isolated RNA from as well as some random ones selected from what I’ve already done
  • Try out/read about using Tri-reagent for isoalting RNA

from Grace’s Lab Notebook

Yaamini’s Notebook: Mixed Effects Models

Assumptions and p-values

Last week, I tackled reviwer comments and used mixed effect models for my analyses. I was right to be concerned about using a glmer with tank as a random effect without any way to assign tanks to some tissues. I reverted back to my binomial GLM. Although I obtained a p-value for my lmer analysis, I wasn’t sure if I was going about it correctly. After some Googling, I found a tutorial from Bodo Winter at UC Merced for linear mixed effect models! It has very clear instructions for constructing mixed effects models with lme4 and for using likelihood ratio tests for p-values.

Likelihood Ratio Tests

In this R script, I modified my code to use likelihood ratio tests instead of an ANOVA workaround. To do this, I needed to build a null model without the fixed effect of interest (either Parental.Treatment or Female.Treatment), and a full model with the effect of interest. I then used an ANOVA to compare both models and get a p-value. A significant p-value means that the effect is part of the most parsimonious model.

When I examined all treatments, I found that parental treatment affected D-hinge counts (Chi-squared = 9.1878, df = 3, p = 0.0269). D-hinge counts were -0.15795 ± 0.10468 lower for low pH female-ambient male and -0.15795 ± 0.10509 lower for low pH female-low pH male. While I have a significant p-value for this model, there is still evidence of a maternal effect. I investigated this further with my second lme using Female.Treatment instead of Parental.Treatment. Female treatment affected D-hinge counts (Chi-squared = 8.1781, df = 1, p = 0.00424). D-hinge counts were -0.21090 ± 0.07033 lower for families with low pH females.


I needed to check the following assumptions for these models:

  1. Normality of residuals
  2. Linearity (because linear mixed effect models)
  3. Homoskedasticity

To check residual normality, I used qqnorm and qqline. Visual inspection did not reveal any obvious violations of the normality assumption, since the data fell on a straight line. I plotted residuals vs. fitted values to check for linearity and homoskedasticity. There was no heteroskedasticity in my residual plots, but I did see striped patterns.


Figure 1. Residual plot for Female.Treatment model.

I’m not sure what this means, and neither Google nor my QSCI 483 notes were of any help. This guide suggests that such a pattern in my residual plot is acceptable, but this Bodo Winter tutorial does not. I might bring it up at lab meeting or ask Tim or Julian for some help. My gut feeling is that there’s some caveat to residual analysis with mixed effects models that I’m missing.

// Please enable JavaScript to view the comments powered by Disqus.


from the responsible grad student

Sam’s Notebook: Mox – Over quota: Olympia oyster genome annotation progress (using Maker 2.31.10)


Well, this is an issue. Checked in on job progress and noticed that we’ve exceeded our storage quota on Mox:


Grace’s Notebook: Skyline DIA with new BLIB made using Walnut

Today Walnut finished up using Walnut to make a new BLIB file with the 2015 C.gigas Oysterseed raw files. Then, I used the new BLIB in SKyline, along with changing the minutes to 5, and double-checking I used the same .mzML files used in Walnut as my results in Skyline. The peaks still look pretty bad. I just emailed Emma…


When I arrived today, there was a “fatal error” in Walnut for February RAW files number 7 and 8. So I tried it again, but it still didn’t work.

So then I went back to owl and re-downloaded the two RAW files. I then converted them to .mzML using MSConvert. Then I added them to Walnut, and it worked! Then I hit “Save BLIB”.

Then, I went through the DIA protocol, but I made a few changes per Emma’s suggestions.

  • I used the new BLIB file I made using Walnut
  • In Step 4e (12), I changed the minutes to 5
  • In Step 4f I made sure I used the same .mzML files as the results that I used to create the BLIB in Walnut img

So, with all this changes, the peaks still looked awful. I didn’t calculate an error rate because it felt like a waste of time. Instead I just looked at about 15 or 20 random peptides and noted that not one of them looked good. There was tons of noise and no clear peaks. I don’t know what to do, so I emailed Emma and sent her a .zip of my Skyline document. I JUST sent it a little bit ago (around 4pm), and I know she’s leaving for a bit, so she may not see it for a while, unfortunately.

I know Nick from Skyline suggested I use the Advanced Peak Picking option, but the protocol doesn’t have super clear instructions for how to do this for DIA Analysis. Emma told me that the pipeline for this isn’t super concrete yet, so that may be why I’m not sure how to do it.

I’ll leave this for now and go back to it on Monday with some fresh eyes, and maybe a response from Emma…

from Grace’s Lab Notebook

Grace’s Notebook: Trying out Walnut for BLIB for 2015 Oysterseed; Crab Pool Update

Today I started using Walnut (upgraded PECAN) to create a new BLIB file for the 2015 Oysterseed project. Hopefully this will improve the error rates in Skyline. Also, I called Pam to work on the NPRB progress report. Additionally, I detail Sam’s updates on the status of the Crab pool samples from his notebook post. The RNA needs some cleaning so he suggests trying RNeasy Cleanup Kit. I will wait until he returns Monday and speak in person with him and Steven to figure out what to do next.


Walnut is an upgraded PECAN and is used to create the BLIB file which I will use in my DIA analysis of the 2015 Oysterseed RAW files in Skyline.

I converted the RAW files to mzML using MSConvert. Then, I uploaded those files to Walnut. Will take a long time so I’ll leave them alone and come back to them tomorrow morning.

NPRB Progress Report

I called Pam this morning to go over the progress report which is due by July 31st (Tuesday). Unfortunately we’ll have to change/add some things to it regarding the issue with the current pools as detailed in my previous notebook post: Crab Pools and Skyline Update. My post is from right before I left for a little vacation to visit family. So, Sam tried to take a closer look at what is going on.

Here’s his post: RNA Cleanup – Tanner Crab RNA Pools

Essentially he used the RNeasy Plus Mini Kit (Qiagen) on the three pools. Then he ran Qubit and Nanodrop1000 with the three pools. The results were similar to what he and I found the day before in that they were not good. The Nanodrop did not detect RNA in the pools. The Qubit didn’t detect RNA in Pool 1, and had very low numbers (~84ng of RNA in each pool) for pools 2 and the MasterPool.

Our target for sequencing is to have 1000ng of RNA in each pool, which total volumes of 50 ul. We are not there.

So, Sam is thinking of potentially having us try using RNeasy cleanup kit on some “test” samples to see if we can get better Qubit readings.

I am gonig to talk about this on Monday with Sam and Steven in person. I will either have Pam call in or update her later. And we’ll have to update the progress report, and I’ll have to do more labwork to try to figure out why these numbers are so low and how we can get these pools to where we need them to be. Stay tuned…

from Grace’s Lab Notebook

DML Analysis: GOterm Update

Sent blastx results and C. virginica genome to Mike Riffle in Genome Sciences. He is going to build me a portal that will do a GOterm enrichment on any set of genes I provide. This is similar to the portal he built for Emma’s geoduck paper.

Linear Mixed Effect Models in R

Good tutorial from Bodo Winter at UC Merced can be found here. He encourages using Likelihood Ratio Tests to obtain p-values.

Yaamini’s Notebook: Mixed Effects Models

lmers and glmers

I got reviewer comments back on my Manchester paper! One of the reviewers suggested I used linear mixed effect models instead of a binomial GLM to evaluate gonad maturity before and after treatment. They also suggested I use a linear mixed effect model instead of a one-way ANOVA to look at differences in D-hinge count too. By using such a model, I can account for random effects, like experimental tank or sire.

To create a linear mixed model in R, I needed the lme4 package. Within the package, I can use lmer for a linear mixed model and glmer for a generalized linear mixed model. The syntax is the same as lm or glm, but I can add random effects with the following term:


Gonad maturation

After talking to Tim and reviewing this guide, I tackled a generalized linear mixed effect model for gonad maturation differences before and after treatment. In my data sheet, I added a column for tank. Pre-treatment oysters were all in the same tank, “Pre,” and post-treatment oysters were coded based on their experimental tank. This is when I ran into a problem. Remember how my tissues got mixed up during processing so I had a bunch of unknown tissues from the ambient treatment? I added all of those to the same “unknown tank.” While they didn’t mix up any treatments, I actually have no way of tracing back the experimental tank for all but two of my ambient treatment tissues. If I cannot assign a tank to these tissues, a glmer that uses tank as a random effect may not be appropriate. I went through with the analysis in this R script, I emailed Tim with my concerns and will update the code as necessary.

D-hinge counts

I did something similar for my D-hinge counts. Instead of a glmer, I used a lmer. First, I specified the sire in a column in this spreadsheet. I used a lmer in this R script and found that the variance for the random effect overlapped zero. This means that the random effect probably doesn’t have much weight, and I can 1) pool by male treatment and 2) use a normal linear model or ANOVA. When looking at the lmer, I had an chi-squared p-value of 0.02445, meaning there were significant differences between all four treatments. I would need to unpack this more to figure out which pairwise differences were significant. I also used a lmer when pooling male treatments, and found that once again, the variance for the random effect (sire) overlapped zero. My ANOVA p-value was 0.00271.

Now that I understand how to use lmer and glmer, I need to understand how to interpret my results. I think a little more reading will help me with this.

// Please enable JavaScript to view the comments powered by Disqus.


from the responsible grad student