search

AlexsLemonade / training-modules

A collection of modules that are combined into 1-5 day workshops on computational topics for the childhood cancer research community.
Other
62 stars 28 forks source link

2024 August issues #798

Open jashapiro opened 2 months ago

jashapiro commented 2 months ago

Issues that came up during the August 2024 training

jashapiro commented 2 months ago

At https://github.com/AlexsLemonade/training-modules/blob/8718d22bb5eb95ea5da06ebc2bf609fffa8d6911/intro-to-R-tidyverse/01-intro_to_base_R.Rmd#L341

The variable is not named example_vector but should be renamed values_1_to_20

jaclyn-taroni commented 2 months ago

Let's keep track of the RRP items on this issue: https://github.com/AlexsLemonade/reproducible-research/issues/184

jaclyn-taroni commented 2 months ago

01 bulk material.

We talk about the other QC report in bulk (i.e., the one not run via the script) as SRR585574, but if one links files as part of the setup, the other reports you get are for SRR585571.

There's also some ambiguity about how we talk about quality trimming since we don't quality trim and the 2020 paper we cite in the next paragraph mentions that trimming appears to have little effect.

The Salmon documentation notes that, given the way we run salmon quant, quantification may be more sensitive to calls that are likely to be erroneous (of low quality) and, therefore, quality trimming may be important.

We can no longer link to --validateMappings in the description of important options of the Salmon docs. However, where the link resolves does currently describe this flag in a note!

jaclyn-taroni commented 2 months ago

We had some issues in 02-gastric_cancer_tximeta with people running the following chunk twice:

https://github.com/AlexsLemonade/training-modules/blob/8718d22bb5eb95ea5da06ebc2bf609fffa8d6911/RNA-seq/02-gastric_cancer_tximeta.Rmd#L125-L129

If we rename what we are assigning to as something like coldata_joined and then we can do:

txi_data <- tximeta(coldata_joined)
sjspielman commented 2 months ago

The pathway analysis notebooks are still using an if statement to create directories. We should update them to use fs::dir_create().

jashapiro commented 2 months ago

We had some issues in 02-gastric_cancer_tximeta with people running the following chunk twice:

https://github.com/AlexsLemonade/training-modules/blob/8718d22bb5eb95ea5da06ebc2bf609fffa8d6911/RNA-seq/02-gastric_cancer_tximeta.Rmd#L125-L129

If we rename what we are assigning to as something like coldata_joined and then we can do:


txi_data <- tximeta(coldata_joined)

Similar potential issue in pathway analysis: Ora

jaclyn-taroni commented 2 months ago

We don't set a seed in the GSEA notebook, but perhaps we should.