Correlating Crab Variables, Fixed

Alright, last time I created a correlation matrix from this table. However, during Pubathon, it was pointed out that I should probably check to make sure I’m doing things correctly. As it turns out, I wasn’t (thanks, Matt!) Basically, I had to separate my normally-distributed and non-normally-distributed variables (as determined by a Shapiro-Wilk test), along with my continuous and categorical variables. However, since all my categorical variables were (unsurprisingly) non-normal, I just was left with two groups – normal and non-normal. I compared the normal ones using Pearson’s test, the non-normal ones with Kendall rank-correlation test, and then cross-compared using Kendall rank-correlation again. This produced 3 sets of plots. Each set contains: 1 dot plot (big dot = higher absolute value of correlation, blue = positive, red = negative) and: 1 correlation chart (number = correlation, plot = plot of correlation) Note: On the following plots, the variables can be…

from Aidan F. Coyle https://ift.tt/3AMj4Ic
via IFTTT

Correlating Crab Variables

A pretty basic question that I haven’t really answered yet is how different variables are actually correlated with each other! So to figure this out, I decided to make a correlation matrix. Each variable applies to one individual crab. So we’re leaving the time dimension out of it, and are just examining covariance between crabs. I made this table a while back. Each row is a value for each individual crab. As a reminder, the three transcriptomes used for alignment in this experiment are: cbai_transcriptomev2.0: unfiltered by taxa cbai_transcriptomev4.0: filtered, presumably only C. bairdi seqs hemat_transcriptomev1.6: filtered, presumably only Hematodinium seqs The following columns are present: Crab ID (A-I) Uniq_ID: This links the notation used in Grace’s analysis (and mine) with some of the earlier Jensen tables Treatment: The temperature treatment the crab was exposed to Timepoints:The number of timepoints (3 for ambient- and lowered-temperature treatment groups, 2 for elevated)…

from Aidan F. Coyle https://ift.tt/3EnumoF
via IFTTT

Adaptive Immune Gene Dive

Previously, I identified several genes with GO terms associated with an adaptive immune response. A writeup to each of these is described in the link, but here’s the suspects: Q92956 (TNR14_HUMAN): Tumor necrosis factor receptor, interacts with CD160 O95971 (BY55_HUMAN): CD160 antigen. Likely plays role in both anti-viral innate immune response along with adaptive immune response P60033 (CD81_HUMAN): CD81 antigen. Whole bunch of roles, inc. structural component, enables receptor complex assembly upon encounter with pathogens, and more. NOTE: P35762 (CD81_MOUSE) was also present in the crab transcriptome. Alright, so we’ve got genes linked to adaptive immunity! This raises two key questions. Are these genes actually related to adaptive immunity? Are these genes correctly identified? Are these genes at a significant expression level? 1 and 2 are the big ones – the presence alone of adaptive immune-related genes would be a neat little finding. But 3 is also important to this…

from Aidan F. Coyle https://ift.tt/3E2mfO0
via IFTTT

Linear Modeling, for real

Longtime readers \ will remember my post from June on using linear modeling to model Hematodinium infection. In case you don’t, I’ll give you a quick recap As part of the QERM 514 class I took, I created a linear model from the survey data that Pam Jensen sent over. These data were Southeast AK Tanner crab surveys from 2007 to 2012, and I focused on modeling Hematodinium infection status. My model found the following linkages to Hematodinium infection. Shell condition: strong linkage. Low for fresh-shell crabs, high for new-shell, then decline as shell age increases Depth: Infection rates higher for crabs in shallower waters Carapace width: Larger crabs were more likely to be infected Julian day: Included in final model, but likely unimportant (p-val = 0.85) Black Mat model: Included in final model, not judged to be significant (p-val = 0.54). However, despite sampling ~800 crabs with Black Mat,…

from Aidan F. Coyle https://ift.tt/3zUI93j
via IFTTT