Skip to content

Latest commit

 

History

History
176 lines (136 loc) · 6.51 KB

bioc-intro.md

File metadata and controls

176 lines (136 loc) · 6.51 KB
Raw

Working notes for the bioc-intro lesson

See Google doc for details.

Defining Target Audience

Minimal questions for bioc-intro course:

Laurent

  • Who are they? Bachelor student (year 2 or 3) in biomedical sciences. No experience in programming or data analysis as part of their curriculum. Some students also struggle with Excel and/or concepts such as file/directory, relative/absolute path. Some haven’t much experience using Excel either.
  • What is the problem they face? At a later stage during their curriculum, they will have to analyse all sorts of data but will never have been taught how to do it or what tools to use.
  • How will your lesson help them address that problem? The bioc-intro lesson should provide them with enough R knowledge to perform simple data manipulation and visualisation to be able to analyse data in a tabular format. In addition, they should also be introduced to the fact that some data won’t fit in a single tabular data, and that omics data typically doesn’t, and that the Bioconductor project offers dedicated tools for such data, as get introduced to basic/important Bioc objects (needed for next courses). They should also be familiar with the concept of package and their installation.

Nitesh

  • who are they? - Students from different backgrounds, Bachelors, Masters, MD, PhD but are looking to gain experience in high throughput genomics analysis with Bioconductor. They are looking to do the analysis themselves and need a place to start. They have domain knowledge in genomics.
  • what is the problem they face? - Data is complex and they have no structured way of analyzing their data. They don’t know what technology or how to use i.e Galaxy, R/Bioconductor, something else….or don’t know what is available.
  • how will your lesson help them address that problem? - Help teach them basics, importing data, analyzing data using Bioconductor packages and data structures, writing results in a coherent reusable manner. (Maybe even “tidy”)

Robert

  • who are they? Researchers generating molecular data without prior computational experience.
  • what is the problem they face? How to read molecular data into R, how to explore that data.
  • how will your lesson help them address that problem? They'll be able to import their molecular data into R, store it into the appropriate Bioconductor data container and explore it with the methods associated with that data container.

Charlotte

  • who are they? PhD student in biology, who has generated some omics data, but never done any analysis on their own. Gets processed data back from the facility.
  • what is the problem they face? They can't explore the results on their own, and are 'stuck' with what is provided.
  • how will your lesson help them address that problem? Get over the initial hurdle of importing and exploring data in R (specifically data stored in 'omics-specific' containers), understanding how to write simple scripts for exploration and plotting of specific aspects of their data.

Jenny

  • who are they? Users of the U of I biotechnology facility (mostly graduate students and post-docs, but a few undergrads and faculty members) who have generated -omics data. They come from a wide range of departments (e.g., Biology, Chemistry, Engineering, Statistics, Computer Science, Agriculture, etc.) so background knowledge varies a lot and they work on a wide range of organisms, often non-model.
  • what is the problem they face? They need to learn how to handle, process and analyze their -omics data.
  • how will your lesson help them address that problem? Show them how to install R and packages, introduce basic R syntax and data structures, teach them how to read and modify code so they can follow and adapt package vignettes.

Summary

Whether undergraduate, graduate or post-docs, the audience isn't expected to have any skills in R. Learners should however be familiar the domain (high-through put biology) in general, and with the measurement of gene expression (without any reference to any technology) in particular.

Objectives and reverse design

Nitesh

  • Objective: Load data from a CSV (excel) file and manipulate it using tidyverse vocabulary in R so that it arranged in a genes as rows and columns as values.

  • Skill: tidyR, dplyr, dyplr verbs, basic R

dat <- read_csv(“mydata.csv”)
dat_1 <- dat %>% select(Gene, Val, metadata) %>% arrange(desc(Val)) %>% pivot_long(...)

Charlotte

  • Objective: Load an expression matrix and sample metadata, select a gene and plot the expression level in different groups.

  • Skills: Read csv file, understand tabular data structure (matrix, data.frame), subset tabular data, merge data from different tables (to combine expression data with sample metadata), ggplot syntax

Laurent

  • Objective: Visualis the expression of a set of genes in different sample (or group of sample).
expression_table %>%
rowname_to_column(gene) %>%
pivot_longer(names_to = “sample”, values_to = “expression”, starts_with(“sample”)) %>%
ggplot(aes(x = sample, y = expression)) +
geom_point() +
facet_wrap(~ gene)
  • Skills needed: get data into R as a data.frame/tibble, understand concept of wide/long tidy data, including familiarity with dplyr, and knowledge of ggplot2 syntax.

Robert

## Import molecular and phenotype data into a Bioconductor data container
library(SummarizedExperiment)
row_data <- read.csv(“expression_data.csv”)
col_data <- read.csv(“phenotype_data.csv”)
se <- SummarizedExperiment(assay=row_data, colData=col_data)
seqdepth <- barplot(colSums(assays(se)$counts))

Jenny

  • Objective: Read in a .gtf or .gff file and assess what feature type and attribute name they should use for alignment and counting.

  • Skills needed: knowledge of the structure of a gff file, how to read one in using rtracklayer, understanding of the structure and accessors of an S4 GRanges object, basic R skills like subsetting via brackets and $, table() and names()

library(rtracklayer)
gff0 <- import("GCF_000146045.2_R64_genomic.gff.gz")
methods(class=GRanges)
table(gff0$type)
gff_exon <- gff0[gff0$type == "exon"]
mcols(gff.exon)
names(mcols(gff.exon))

Summary

The goal of this course is to bring newcomers, with no prior experience of R, to be able to perform standard data science tasks and visualisation using the tidyverse and familiarise themselves with other concepts so that they now feel confident to start learning Bioconductor with the next courses.