Post with little updates on what I did last week!

Summer 2021 Work

Did some more PCAs, and DESeq2 analyses - nothing is annotated yet, though, so this is just looking into differences between treatment groups and disease sign groups preliminarily.

Note:
For summer 2021 work, we did a bunch of experiments, but I was only able to get enough RNA for sequencing for some samples.

I have RNAseq data for samples that are part of Experiment A and Experiment B. Datasheet for all library metadata from Summer 2021: grace-ac/paper-pycno-sswd-2021/data/sample_metadata.csv.

The count matrix for all 32 libraries lives here: grace-ac/paper-pycno-sswd-2021/data/kallisto_count_matrix_rounded.tab. The counts were derived using kallisto (code), and using the gene list FASTA from the published Pycnopodia helianthiodes genome.

Experiment A began on 9/23/2021. Stars were injected with either 0.45micron filtered live tissue homogenate from a wasting star (microbial group), heat-killed tissue homogenate from a wasting star (control group), or unfiltered live tissue homogenate from a wasting star (exposed group). Stars were observed 2x daily and sampled at disease sign along with control pair. The RNAseq data timeline control and exposed groups are not balanced due to lack of sufficient RNA for sequencing for some samples. Comparisons made for those libraries are shared below under the title “Experiment A”.

Experiment B began on 10/05/2021. Stars were injected with heat-killed tissue homogenate from a wasting star (control group) or unfiltered live tissue homogenate from a wasting star (exposed group). Some of the control stars in this experiment were previously injected with the 0.45micron filtered live tissue homogenate in Experiment A but never showed any disease signs, and so they became part of the control group. They continued to remain disease-sign free throughout the remainder of Experiment B. Comparisons made between libraries from Experiment B are shared below under the title “Experiment B”.

Experiment A:

library_ID star_ID previous_exposure treatment sample_date experiment_day disease_sign
PSC.19 H16 tank_exposure na 9/23/21 na armdrop
PSC.23 H15 tank_exposure control 9/26/21 3 armdrop
PSC.24 E13 tank_exposure 0.45live 9/27/21 4 armdrop
PSC.34 H12 tank_exposure 0.45live 10/2/21 9 armtwist
PSC.35 E05 na unfilteredlive 10/2/21 9 armdrop
PSC.36 E04 na unfilteredlive 10/2/21 9 armdrop
PSC.37 E01 na unfilteredlive 10/2/21 9 armdrop
PSC.38 E12 na 0.45live 10/2/21 9 na
PSC.39 H05 na control 10/2/21 9 na
PSC.40 H04 na control 10/2/21 9 na
PSC.42 H03 na control 10/3/21 9 na
PSC.43 E02 na unfilteredlive 10/3/21 10 armdrop
PSC.48 E14 na 0.45live 10/7/21 12 na
PSC.49 E16 na 0.45live 10/7/21 12 na

Armdrop Vs. Healthy

Library PSC.19 is from a star that had a shared exposure in a co-housing tank situation with PSC.23, PSc.24, and PSC.34… it did not survive to make it to Experiment A, but because it has the same exposure history as those three libraries that WERE included in Experiment A… I perhaps weirdly decided to include this library in the below preliminary comparison.

img

PERMANOVA: 0.001. PCA showing significant difference in RNAseq data between stars that were sampled at arm drop disease sign compared to stars that were healthy at sampling.

Note: PERMANOVA performed in R code: grace-ac/paper-pycno-sswd-2021/code/04-PCAplots.Rmd, lines 240-274.

img

Volcano plot of the 4,093 DEGs found using DESeq2 when comparing the libraries shown above in the PCA.

Control Vs. Unfiltered Live

img

PERMANOVA: 0.222 PCA showing that there is no difference in RNAseq data between stars in treatment of heat-killed tissue homogenate compared to stars injected with unfiltered live tissue homogenate. I don’t think this result is due to the treatment groups, but rather due to their exposure history where PSC.23 was cohoused with the tank outbreak.

Note: PERMANVA performed in R code: grace-ac/paper-pycno-sswd-2021/code/04-PCAplots.Rmd, lines 319-348.

img

Volcano plot of the 36 DEGs found using DESeq2 when comparing the libraries by treatment group as shown in the above PCA.

Control Vs. 0.45 micron filtered (microbial fraction)

img

PERMANOVA: 0.846.
PCA showing that there is no significant difference in RNAseq data between libraries that are from stars treated with a 0.45 micron filtered live tissue homogenate and the control group (heat-killed tissue homogenate).

img
Volcano plot of the 4 DEGs found using DESeq2 when comparing the libraries by treatment group as shown in above PCA.

Unfiltered live sick tissue homogenate Vs. 0.45 micron filtered sick tissue homogenate (microbial fraction)

img

PERMANOVA: 0.367.
PCA showing that there is no significant differences in RNAseq data between libraries from stars injected with unfiltered tissue homogenate from a sick star when compared to stars injected with 0.45 micron filtered tissue homogenate from a sick star. Note: I don’t know that the treatment groups are driving these results, but more likely the fact that TWO of the five stars in the 0.45 micron treatment were cohoused in an accidental exposure event to a wasting star from the field.

img

No significant DEGs were found. Again, I think this is due to the cohousing tank exposure that occurred for 2/5 of the 0.45 micron filtered treatment group.

Take-aways from Experiment A Comparisons:

I think the previous tank exposure that occurred for several of these stars makes it so that we can’t say much about how the experimental treatment impacted their gene expression. But, I do think that comparing based on disease sign and how that gene expression compares to Experiment B could be interesting.

Experiment B:

library_ID star_ID previous_exposure treatment sample_date experiment_day disease_sign
PSC.56 E06 0.45micronLive control 10/14/21 9 NA
PSC.81 E07 0.45micronLive control 10/18/21 13 NA
PSC.61 E09 0.45micronLive control 10/15/21 10 NA
PSC.76 E10 0.45micronLive control 10/17/21 12 NA
PSC.52 E11 0.45micronLive control 10/14/21 9 NA
PSC.54 E18 0.45micronLive control 10/14/21 9 NA
PSC.63 H03 NA control 10/15/21 10 NA
PSC.73 H03 NA control 10/16/21 11 NA
PSC.58 H05 NA control 10/14/21 9 NA
PSC.64 H05 NA control 10/15/21 10 NA
PSC.59 H01 NA unfilteredlive 10/14/21 9 armtwist
PSC.57 H06 NA unfilteredlive 10/14/21 9 armdrop
PSC.67 H09 NA unfilteredlive 10/15/21 10 armdrop
PSC.69 H04 NA unfilteredlive 10/15/21 10 armdrop
PSC.71 H18 NA unfilteredlive 10/15/21 10 armdrop
PSC.75 H08 NA unfilteredlive 10/16/21 11 armdrop
PSC.78 H10 NA unfilteredlive 10/17/21 12 armdrop
PSC.83 H07 NA unfilteredlive 10/18/21 13 armcross

Control Vs. Exposed

Stars H03 and H05 are controls that were samples twice, each, so I just selected one of the libraries to include in the below PCA and volcano plots.

img

PERMANOVA: 0.001. PCA showing that there is a significant difference in RNAseq data between control and exposed treatment groups.

Note: PERMANOVA performed in R code: grace-ac/paper-pycno-sswd-2021/code/04-PCAplots.Rmd, lines 75-109.

img

Volcano plot of the 7,834 DEGs found using DESeq2 comparing the control vs exposed treatment groups as shown in the above PCA.

Summer 2023 Multi-Species Planning

I’ll be extracting RNA from samples from the 6 exposed bins, each with a wild juvenile Pycnopodia helianthoides, wild Pisaster ochraceus and wild Dermasterias imbricata, from Day 12 of the multi-species experiment. That will be n=18 samples that will (provided there’s enough RNA) be sent of for total RNA-sequencing.

This time point was selected because it will be a good starting point that will capture a lot of information that will help answer multiple questions.

Summer 2023 Image Organization

I finally finished uploading, labelling, and organizing all the photos I took of the stars from Summer 2023 experiments! Woo!

Other

Starting the process of re-vamping my website/notebook! Making it pretty, updating project details, links, etc.