BIOSCI 220 Notes (Week 01–02): R, Tidyverse, and Indigenous Data Sovereignty

R and RStudio basics

• What is R vs what is RStudio
  • R: the computing language for statistics and data analysis.
  • RStudio: a user-friendly integrated development environment (IDE) that helps you write, run, and organize R code.
  • Learning focus: how to use R and how RStudio facilitates workflow (console, script editor, plots, etc.).
• Why use R? (Slide content overview)
  • Free and open source.
  • Huge online support network.
  • Flexible; if you can code it, you can do it.
  • Large community and ecosystem of packages.
  • Relevance to research data analysis and reproducible workflows.
• Why use RStudio? (Slide content overview)
  • Works nicely with R; integrates multiple tools in one interface.
  • Large online support network and extensibility.
  • Distinctive features: project management, built-in plots, help panes, etc.
• Key terminology (from the transcript):
  • Functions: R commands that tell your computer what to do.
  • Arguments: inputs you provide to functions that influence outputs.
  • Packages: collections of functions; extend R’s capabilities.
  • Running code: issuing commands in the console or running a script.
  • Objects: values stored in R (variables, data frames, etc.).
  • Script: a text file containing a set of commands and comments.
  • Comments: notes within a script to document what’s happening.
• Example R code snippets mentioned in the slides (illustrative):
  • Installing packages and loading libraries:
  • $install.packages(c(\"dplyr\", \"remotes\"))$
  • $library(dplyr)$
  • Using GitHub installations (as shown):
  • $remotes::install_github(\"djnavarro/jasmines\")$
  • Simple art-style code snippet (illustrative):
  • $start<em>seed(123) \%\% entity</em>heart(grain = 400) \%\% unfold<em>tempest(iterations = 10) \%\% style</em>ribbon(background = \"wheat\")$
  • Note: visuals and art-focused R packages are shown as examples of R’s flexibility.

R data types

• Integer (int): whole numbers; examples include $1, 0, 220$; class is labeled as $"integer"$ or $int$ in R.
• Numeric (num / dbl): numbers with decimals or fractions; examples include $-56.94, 1.3$; class is $"numeric"$ or $dbl$ in R.
• Logical (lgl): TRUE or FALSE; class is $"logical"$.
• Character (chr): text strings; examples include $"Charlotte"$ and $"BIOSCI220"$; class is $"character"$.
• Quick summary (from slide content):
  • Integers: $ext{class} = ext{integer (int)}$.
  • Numeric: $ext{class} = ext{numeric (num, dbl)}$.
  • Logical: $ext{class} = ext{logical (lgl)}$.
  • Character: $ext{class} = ext{character (chr)}$.

Tidy data and data structure fundamentals

• Tidy data (Hadley Wickham): a standard for mapping dataset meaning to structure.
  • Quote: "Tidy datasets are easy to manipulate, model and visualize …" ( Hadley Wickham, Tidy Data, JSS 2014 ).
• Core tidy data rules:
  1) Each variable forms a column.
  2) Each observation forms a row.
  3) Each value is a single cell.
• Why tidy data matters:
  • Facilitates data manipulation, modeling, and visualization with a small, consistent toolbox.
• Practical takeaway:
  • In tidy data, you can reliably apply the same functions across many datasets without special handling for messy shapes.

Piping and data wrangling with tidyverse

• Piping operator: %>%
  • Concept: chaining operations in a readable left-to-right sequence.
  • Example pattern: data %>% op1() %>% op2() %>% op3()
  • A concrete example from slides:
  • $pau a \%\% group<em>by(Species) \%\% summarise(av</em>length = mean(Length))$
• Common errors (typical beginner mistakes):
  • Typing $library(tidy)$ instead of the package name; error: "there is no package called 'tidy'".
  • Typing $lbrary(tidyverse)$ (typo) resulting in: "could not find function 'lbrary'".
  • Attempting to use a function that doesn’t exist yet, e.g., $read_csv$ without loading the readr package or the tidyverse.
• Example data wrangling workflow with Palmer Penguins (tidyverse):
  • Start by loading libraries:
  • $library(palmerpenguins)$
  • $library(tidyverse)$
  • Remove missing values:
  • penguins o{}nafree
    <- penguins \%\% drop_na()
  • Basic grouped summaries:
  • $penguins o{}nafree \%\% group<em>by(species) \%\% summarise(av</em>length = mean(bill<em>length</em>mm))$
  • Example of a more complex pipeline:
  • $penguins o{}nafree \%\% filter( sex != \"male\") \%\% select(c(\"species\", \"island\", \"body<em>mass</em>g\")) \%\% group<em>by(species, island) \%\% summarise(total</em>mass<em>g = sum(body</em>mass<em>g)) \%\% pivot</em>wider(names<em>from = c(island), values</em>from = total<em>mass</em>g)$
• Summary from data wrangling recap (Page 27):
  • A concise, reproducible pipeline using library(palmerpenguins) and library(tidyverse).
  • Demonstrates filtering, selecting, grouping, summarising, and reshaping with pivot_wider.

Working with Palmer Penguins data (exploratory data analysis)

• Demonstration outputs shown in slides included:
  • Example: penguinsnafree %>% groupby(species) %>% summarise(avgeragebilllength = mean(billlengthmm))
  • Result (illustrative):
  • species: Adelie, Chinstrap, Gentoo
  • avgeragebilllength: $38.8, 48.8, 47.6$ (units: mm)
• Example of a grouped summary table:
  • Output shows per-species averages, e.g.,
  • $ext{species} = \text{Adelie} \rightarrow 38.8$
  • $ext{Chinstrap} = 48.8, ext{Gentoo} = 47.6$

Data wrangling with real datasets and expressions

• Example: filtering, selecting, grouping, and summarising to compute total body mass by species and island, then pivotting to wide format for comparison:
  • Pipeline (cleaned version):
  • $penguins o{}nafree \%\% filter(., sex != \"male\") \%\% select(c(\"species\",\"island\",\"body<em>mass</em>g\")) \%\% group<em>by(species, island) \%\% summarise(total</em>mass<em>g = sum(body</em>mass<em>g)) \%\% pivot</em>wider(names<em>from = c(island), values</em>from = total<em>mass</em>g)$
• Interpretation:
  • The resulting table shows, for each species, the total body mass across different islands.

ggplot2 basics and data visualization concepts

• ggplot2 overview: a package in the tidyverse for creating statistical graphics.
• Three key components of every ggplot:
  1) Data: the data object (e.g., penguinsnafree). 2) Aesthetic mappings: how variables map to visual properties (e.g., x, y, color). 3) Layers: render observations with geoms (e.g., geompoint, geom_bar).
• A minimal example (clean, correct form):
  • $library(palmerpenguins)$
  • $library(tidyverse)$
  • $penguins o{}nafree \%\% ggplot(data = penguins o{}nafree, aes(x = bill<em>length</em>mm, y = bill<em>depth</em>mm, color = species)) + geom_point(size = 2)$
• Customizing colors and axes:
  • Example: $+ scale<em>color</em>manual(values = c(\"darkorange\", \"darkorchid\", \"cyan4\"))$
  • Example axis labels: $+ xlab(\"Bill length (mm)\") + ylab(\"Bill depth (mm)\")$
• Common ggplot2 plots shown:
  • Scatter plots: billlengthmm vs billdepthmm by species.
  • Bar plots with facetting by island: $ggplot(…) + geom<em>bar(…) + facet</em>wrap(~island)$
• Practice reminder: a plot typically requires data, aesthetics, and at least one geom layer.
• Ethics and impact of visualization (key notes):
  • Visualizations are not neutral; they influence interpretation and decision-making.
  • Quotes from practice slides emphasize responsibility in how data patterns and conclusions are presented (e.g., Correll, Rougier et al.).
  • Ethical data practices in visualization require clarity, honesty, and awareness of audience and biases.

Data visualization ethics and interpretation

• Direct quotes and implications:
  • "We have obligations in that we have a great deal of power over how people ultimately make use of data, both in the patterns they see and the conclusions they draw." (Michael Correll)
  • General guidance: choose plots and scales that accurately reflect the data without overstating findings.
• Key takeaway: Data visualization is a powerful tool; responsible use is essential for credible science.

Indigenous Data Sovereignty (IDSov) and related frameworks

• What is Indigenous Data Sovereignty (IDSov)?
  • Indigenous Data (ID): digital information derived from Indigenous people, their language, culture, and resources.
  • IDSov: the inherent rights and interests Indigenous peoples have over the collection, ownership, and application of their data.
  • Foundational reference: Kukutai & Taylor (2016) and Māori data sovereignty resources.
• International and national rights landscape:
  • UN Declaration on the Rights of Indigenous Peoples (UNDRIP), 2007 (UN General Assembly resolution A/RES/61/295).
  • Convention on Biological Diversity (CBD), 1992.
  • Nagoya Protocol (2010): fair and equitable access and sharing of benefits arising from genetic resources.
  • New Zealand’s Te Tiriti o Waitangi (Treaty) and alignment with UNDRIP principles.
• UNDRIP key themes (UNDRIP has 46 articles):
  • Self-determination, equality and non-discrimination, participation, culture, land and resources.
  • Article 31 explicitly discusses Indigenous rights to maintain, control, protect, and develop cultural heritage, traditional knowledge, and traditional cultural expressions (including genetic resources and knowledge).
• Te Tiriti o Waitangi (NZ Treaty) alignment with human rights:
  • Article 1: right to self-determination for incoming settlers; democratic and citizenship rights.
  • Article 2: right to self-determination for Tangata Whenua; indigenous rights and property rights.
  • Article 3: equality and non-discrimination.
  • Article 4: freedom of religion and beliefs.
  • These articles align with international human rights documents like UNDRIP.
• CARE Principles of Indigenous Data Governance (wilkinson et al.; GIDA): four interrelated principles for data governance:
  • Collective Benefit (C): data ecosystems should enable Indigenous peoples to derive benefits.
  • C1. Inclusive development and innovation.
  • C2. Improved governance and citizen engagement.
  • C3. Equitable outcomes; benefits should accrue to Indigenous communities through responsible use and benefit-sharing.
  • Authority to Control (A): Indigenous peoples’ rights to control data must be recognized and respected.
  • A1. Recognizing rights and interests; informed consent and governance of collection.
  • A2. Data for governance; data must be accessible to empower self-determination.
  • A3. Governance of data; Indigenous communities should lead stewardship and access decisions.
  • Responsibility (R): responsible data use, sharing outcomes, and governance.
  • R1. Positive relationships; data use builds on trust and respect.
  • R2. Expanding capability and capacity; developing Indigenous data workforce and infrastructure.
  • R3. Indigenous languages and worldviews; data grounded in languages and lived experiences.
  • Ethics (E): minimizing harm, ensuring justice, and considering future use.
  • E1. Minimizing harm and maximizing benefit within Indigenous ethical frameworks and UNDRIP rights.
  • E2. Justice; address power/resource imbalances.
  • E3. Future use; consider potential future harms and benefits.
• Ngā Tikanga Paihere (Tikanga framework) – the 5 Safes framework with ethical data use practices:
  • Principle 1: Have appropriate expertise, skills, and relationships with communities.
  • Pūkenga (skills) and Whakapapa (community relationships).
  • Principle 2: Maintain public confidence and trust; Pono (accountability) and Tika (transparency and value).
  • Principle 3: Use good data standards and practices; Wānanga (organisational practices) and Kaitiaki (data stewardship).
  • Principle 4: Have clear purpose and action; Wairua (community good) and Mauri (data provenance).
  • Principle 5: Balance benefits and risks; Tapu (sensitivity) and Noa (benefit and opportunity).
• Core ethical takeaway from IDSov slides:
  • Maori data sovereignty is about collective rights and interests; Māori data governance refers to mechanisms that give effect to those rights.
  • The phrase “Not about us, without us” emphasizes inclusive decision-making and co-creation.

Indigenous data sovereignty in practice: case studies (New Zealand)

• Aotearoa Variome Project (Otago):
  • Goal: co-develop a catalog of genetic variants for New Zealanders to understand population variation.
  • Governance: Māori-led with co-development of consenting processes and ensuring benefits align with Te Ao Māori and kaitiakitanga.
  • Outcome: a Māori-led genomic catalogue with governance designed to reflect Indigenous values.
• Rakeiora Project (Auckland):
  • A pilot precision medicine infrastructure platform.
  • Stand-alone scalable computational platform with co-design, governance, data sovereignty, and co-innovation incorporating mātauranga whakapapa.
  • Emphasis on co-development between Māori and non-Māori leaders to ensure equitable access and governance.
• Summary of practical implications:
  • National (Te Tiriti o Waitangi) and international (UNDRIP) instruments support Indigenous rights over genetic/genomic data.
  • Ethical frameworks and governance structures (CARE, Ngā Tikanga Paihere) guide responsible data use and equitable benefit sharing.
  • The goal is to expand genomic technologies and ethical data practices without exacerbating health disparities.

Diversity in genomics research and the GWAS data landscape

• Problem statement: Historically, genomic studies have underrepresented Indigenous and non-European populations.
• Key statistics (illustrative, pulled from slides):
  • Total GWAS studies: $373$; samples: $1.7 ext{ million}$ in one period, $2{,}511$ studies with $35 ext{ million}$ samples, and $3{,}639$ studies with $40 ext{ million}$ samples (illustrative tabular data).
  • Ancestry composition over time (examples):
  • European Ancestry vs Non-European Ancestry in early waves: 88 ext{%} vs 12 ext{%} (2005–2018).
  • 2005–2009: 96 ext{%} European vs 4 ext{%} non-European.
  • 2005–2016: 80 ext{%} European vs 20 ext{%} non-European.
  • Indigenous ancestry contribution to GWAS samples: < 1 ext{%} (very small share; often reported as micro-percentage values like 0.06 ext{%}, 0.05 ext{%}, 0.02 ext{%} in various periods).
• Diversity monitor (global context): interactive dashboards showing ancestry by participants over time, with parent terms like European, Asian, African, and Indigenous ancestry. These dashboards illustrate the persistent lack of representation for Indigenous and non-European groups in GWAS.
• Implications:
  • Limited diversity in GWAS reduces the generalizability of genetic findings to all populations.
  • There is a need for deliberate efforts to diversify genomic databases to improve health equity.

Global initiatives to diversify genomic data and ethical considerations

• Global initiatives aim to broaden representation and governance in genomic data:
  • Silent Genomes Project: aims to diversify genomes and improve accessibility and governance.
  • A draft human pangenome reference (Nature 2023): broadening the reference to represent global diversity beyond a single reference genome.
• Guiding question: Who benefits from this research and how should next-generation genomics tools be used to avoid widening health disparities?
• Practical takeaways:
  • Building inclusive genetic resources requires collaboration with Indigenous communities and alignment with IDSov principles.
  • Policies and infrastructures should support equitable access, governance, and benefit-sharing.
• Quote (contextual): Dr. Eric Green (NHGRI): to realize the promise of genomic medicine, organizations must commit to equitable expansion of responsible genomic data use.

Summary of IDSov frameworks and practical implications

• IDSov core definitions:
  • Indigenous Data (ID): digital information derived from Indigenous groups.
  • Indigenous Data Sovereignty (IDSov): inherent rights/claims over collection, ownership, and use of Indigenous data.
• Core principles and frameworks:
  • CARE Principles: Collective Benefit, Authority to Control, Responsibility, Ethics.
  • Ngā Tikanga Paihere: a Tikanga-based framework emphasizing the 5 Safes and related ethical practices.
  • The CARE and Tikanga frameworks together provide a culturally grounded approach to data governance that foregrounds Indigenous rights, participation, and benefits.
• Key outcomes:
  • Data sovereignty supports self-determination and governance over data that concern Indigenous communities.
  • Research should be designed to maximize collective benefits and minimize harms, with transparent governance and reporting.
• Takeaway phrase: “Not about us, without us” signals the necessity of Indigenous involvement in all stages of data collection, analysis, and dissemination.

Case studies and practical implications recap

• Aotearoa Variome Project:
  • Maori-led governance with co-design of processes to ensure consent is culturally appropriate and benefits align with Te Ao Māori and kaitiakitanga.
  • Produces a catalog of genetic variants representative of New Zealanders while upholding data sovereignty.
• Rakeiora Project:
  • Pathfinding genomic medicine infrastructure in Aotearoa/New Zealand.
  • Stand-alone platform with co-design and governance; incorporates mātauranga whakapapa; co-innovation among Māori and non-Māori leadership.
• Overall implication:
  • National and international agreements support Indigenous rights in genetic data.
  • Ethical frameworks promote equitable benefits and prevent health disparities from widening as genomic medicine expands.

Conceptual connections and practical implications for exam readiness

• Core ideas to connect across modules:
  • R basics and tidyverse underpin modern data science workflows used in ecology and genomics.
  • Tidy data and the pipe operator enable clean, reproducible data manipulation pipelines.
  • ggplot2 provides a consistent approach to data visualization; ethical visualization emphasizes accurate and responsible representation of data.
  • Indigenous Data Sovereignty introduces critical, ethics-centered considerations for research involving Indigenous data, highlighting power dynamics, governance, and the need for community-led data stewardship.
• Practical exam-ready takeaways:
  • Be able to explain the three core components of a ggplot2 plot and to assemble a basic scatter plot with proper aesthetics.
  • Describe the tidy data principles and give examples of how a messy dataset would be transformed into tidy form using a pipe chain.
  • Define IDSov and CARE principles, and explain how they would guide a hypothetical genomics study involving Indigenous populations.
  • Summarize the international agreements (UNDRIP, CBD, Nagoya) and Te Tiriti o Waitangi, and discuss how they intersect with data governance and genomic research in a real-world setting.
• Notable formulas or equations in these slides:
  • There are few explicit mathematical formulas in the deck; the emphasis is on data manipulation, summaries, and visualizations. When summarising data, common expressions include:
  • $av ext{-}length = \text{mean}(bill_length_mm)$
  • $total_mass_g = \sum(body_mass_g)$
  • In ggplot2 contexts, expressions are typically of the form: $\text{ggplot}(data, aes(x, y, colour)) + geom_point()$
• Final reflection:
  • The course emphasizes both technical data skills (R, tidyverse, ggplot2) and ethical, sociocultural considerations (IDSov, CARE, Tikanga) that shape how data-driven insights are produced and shared.

Quick glossary (key terms from the slides)

• R: The programming language used for statistics and data analysis.
• RStudio: Integrated development environment for R.
• tidy data: A dataset where each variable is a column, each observation is a row, and each value is a single cell.
• pipe (%>%): Operator enabling readable chains of data transformations.
• ggplot2: Grammar of graphics package for building layered plots.
• Palmer Penguins: A dataset frequently used in tutorials; example data for teaching data wrangling and visualization in R.
• Indigenous Data Sovereignty (IDSov): Inherent rights and interests of Indigenous peoples in relation to their data.
• CARE Principles: Collective Benefit, Authority to Control, Responsibility, Ethics (data governance framework).
• Ngā Tikanga Paihere: Tikanga-based ethical framework with the 5 Safes for responsible data use.
• UNDRIP: United Nations Declaration on the Rights of Indigenous Peoples (2007).
• CBD and Nagoya Protocol: Frameworks related to access, benefit-sharing, and biodiversity.
• Te Tiriti o Waitangi: The Treaty principles guiding Indigenous (Māori) rights in Aotearoa/New Zealand.
• Aotearoa Variome Project: Maori-led catalog of genetic variants for New Zealand.
• Rakeiora Project: Genomic medicine platform in NZ with co-design and data sovereignty.