CURRENTLY UNDER CONSTRUCTION

Semester: Fall 2026

Meeting time: Mon/Fri - 10:10-11:55am, Wed - 9:05-9:55am

Location:

Mon/Fri: CDS B62

Wed: SAR103

Zoom: Posted on Blackboard

Office hours: By appointment — contact information on Blackboard

Joey Wednesdays, 10-11am LSEB 101

Monday, 3-4pm LSEB 101

Contents

Course Objectives

  • Learn the molecular mechanisms and basic data analysis steps that underlie common next-generation sequencing experiments
  • Develop proficiency in creating bioinformatics workflows with an emphasis on reproducibility and portability
  • Gain experience generating and interpreting bioinformatics analyses in a biological context

Topics covered include:

  • High Throughput Sequencing Technologies (RNAseq, ChIPseq, scRNAseq) and various omics technologies (Proteomics, Metabolomics, etc.)
  • Computational Workflow Tools (Snakemake, Nextflow)
  • Reproducibility and Replicability Tools (Git, Docker, Conda)
  • Bioinformatics Databases and File Formats
  • Responsible use of LLMs

Course Description

This course covers modern bioinformatics with a specific focus on the analysis of next generation sequencing data. Lectures cover a mix of biological and computational topics necessary for the technical and conceptual understanding of current high-throughput genomics techniques, including the molecular mechanisms of the assays, basic data analysis workflows, and translating results into biological conclusions.

Students build computational workflows that perform end-to-end analyses of sequencing data from RNA-sequencing, ChIP-sequencing, and single-cell RNA-sequencing experiments. The course emphasizes reproducibility and portability throughout. AI tools are treated as part of that workflow: used deliberately, evaluated critically, and documented with the same rigor as any other methodological choice.

Labs focus on practical activities with the tools and technologies needed to analyze and interpret sequencing data.

Prerequisites

Basic understanding of biology and genomics. Any of these courses are adequate prerequisites: BF527, BE505/BE605. Students should have some experience programming in a modern language (R, Python, C, Java, etc.).

Working familiarity with Git and the command line is strongly recommended.

Required Software

All you need is a laptop. Course computing runs on BU’s Shared Computing Cluster (SCC), accessed via a browser-based VSCode session — no local installation of bioinformatics tools is required. You will be automatically provided with access to the shared computing cluster.

If you do not currently have a GitHub account, please make one prior to the start of the first class. It is free and one of the most widely used platforms for hosting git repositories.

Instructor and TAs

Joey Orofino

Contact information available on Blackboard

As instructor, I will:

  1. Learn and correctly pronounce everyone’s preferred name/nickname
  2. Use preferred pronouns for those who wish to indicate this to me
  3. Work to accommodate language-related challenges (I will do my best to avoid idioms and slang)

Course Values and Policies

Respect. Every background, race, color, creed, religion, ethnic origin, age, sex, sexual orientation, gender identity, and nationality is welcome in this course. Disrespectful language, discrimination, or harassment of any kind are not tolerated and may result in removal from class or the University. Incidents can be reported to the instructor, the Bioinformatics Program leadership, or the BU Equal Opportunity Office.

Collaboration. Collaboration is encouraged. You may work with others, share ideas and code, and use any resources available to you — including the internet. Your written reports must reflect your own analysis, interpretation, and understanding of the results.

Attendance. Lab attendance is tracked through Git commits. Each lab session has associated tasks that should be committed to your repository during or shortly after the session. Regular commit activity is expected and counts toward the 20% participation grade.

AI and LLM tools. AI tools (ChatGPT, GitHub Copilot, Claude, etc.) are a legitimate and expected part of this course. Used critically, transparently, and with appropriate skepticism, they will be a significant asset to your skillset in this course and beyond. Each assignment has an explicit AI use level that tells you what is permitted and why. See the AI Use in This Course section for the full framework.

Flexibility. If something comes up that affects your ability to participate, let me know and we’ll work it out. You don’t need to share details you’re not comfortable sharing. BU Student Health Services is available if you need additional support.

Projects Overview

Each project asks you to build a Nextflow pipeline that performs an end-to-end analysis of a real sequencing dataset, then write up the results as sections of a scientific publication. Projects increase in complexity and decrease in scaffolding as the semester progresses.

  • Project 1 — Genome Assembly: Assemble a bacterial genome from long reads, assess assembly quality, and annotate predicted genes.
  • Project 2 — RNA-seq: Quantify gene expression from paired-end RNA-seq data, identify differentially expressed genes, and interpret the results in a biological context.
  • Project 3 — ChIP-seq: Call transcription factor binding peaks, perform motif enrichment analysis, and compare binding profiles across conditions.
  • Final Project: An open-ended analysis of a dataset and question of your choosing, integrating methods from across the course.

All pipelines are built with Nextflow and run on the SCC using Singularity containers, with results version-controlled in Git.

A note on lab numbering: Labs are numbered by topic rather than chronological order. Some labs appear in the schedule out of numerical sequence — each number is a stable topic reference, not a position indicator.

AI Use in This Course

This course uses the AI Assessment Scale (AIAS) (Perkins, Furze, Roe and MacVaugh, 2024) to make AI policy transparent and consistent across all assignments. The AIAS describes five levels of AI involvement in assessment, from no AI to full AI-human collaboration. Rather than a single blanket policy, each component of this course has an assigned level that reflects its learning goals. The level tells you both what level of use is expected and why that choice was made.

The goal is to use AI in ways that serve your learning rather than substitute for it.

AIAS Levels at a Glance

Level Name What it means
1 No AI Completed entirely without AI assistance
2 AI Planning AI may support brainstorming and structuring; not present in final submission
3 AI Collaboration AI assists in developing and refining work; human judgment directs and evaluates output
4 Full AI AI used throughout; all AI-generated content must be cited and critically evaluated
5 AI Exploration Open-ended co-design with AI; boundaries defined collaboratively

For Level 4 work, AI use is structured around test-driven development — specifications and verification plans come before prompting. Details are in the Final Project section below.

Assignments and Their Levels

Labs — Level 1–3 Labs are participation exercises, not graded assignments. Each lab has an assigned AI level that reflects its goals. When appropriate, you are encouraged to use AI tools freely to work through the material. Pushing changes to your repo is how participation is tracked.

Projects 1 & 2 — Level 2 (AI Planning) These early projects ask you to build foundational skills by working through pipelines largely by hand. You may use AI for brainstorming, planning, and debugging small well-defined subtasks but the pipeline logic, parameter choices, and written analysis should be yours. The idea is to develop confidence and the ability to evaluate an analysis before leaning on tools that can do it for you. Developing and running a pipeline you’ve built yourself, whose parameters you have chosen and whose outputs you have questioned, will give you the intuition to recognize when an LLM-generated version is subtly wrong. Importantly, it also builds the vocabulary to ask for the right thing in the first place by generating precise, well-grounded prompts stemming from an understanding of the problem to be solved.

Each submission includes a trust map: a table of every pipeline step annotated with a trust level (high / medium / low) and a rationale. At this stage, most of your pipeline is hand-built, so the trust map is primarily an exercise in articulating why you believe each step is correct. “It ran without errors” is not a rationale.

Projects 3 & 4 — Level 3 (AI Collaboration) By this point you have enough hands-on experience to use AI as a genuine collaborator. You may use AI tools more broadly for drafting code, exploring methods, and structuring your write-up but your critical evaluation of the output is the work. Each submission includes a trust map and verification evidence: for every step you flag as medium or low trust, a specific verification method carried out and documented. Not “I will check the output” but “I confirmed the number of significantly DE genes is plausible for this experimental design, and I verified the fold change direction for three known marker genes against a primary database.”

Final Project — Level 4 (Full AI) AI use is not just permitted here; it is expected. The trust map and verification practice from earlier projects now governs the full AI-assisted workflow. Note that Level 4 applies to the project as a whole; individual components carry their own levels. The specification document and trust map are Level 1. They must be completed without AI assistance. Their value depends entirely on representing your own thinking before any AI interaction occurs.

Your submission has four components:

1. Specification document — Level 1 (No AI) Written before any AI interaction. For each step in your pipeline, describe what a correct output looks like: its format, expected value ranges, and any properties you could check programmatically. This is your test suite. Writing it first forces you to understand the analysis before you ask AI to help build it, and gives you a concrete standard against which to evaluate what comes back.

2. Trust map — Level 1 (No AI) A table of every pipeline step annotated with a trust level (high / medium / low) and a rationale. The trust level reflects how likely AI-generated output is to be subtly wrong at that step, and why. A step with a high trust level still needs a rationale and “it looked right” is not one. Steps involving experimental design choices, biological interpretation, or tool flags that depend on your specific data are almost always low trust.

3. Verification evidence — Level 3 (AI Collaboration) For every step flagged as medium or low trust, a specific verification method carried out and documented. Unit tests, manual spot checks, and cross-references against published results all count.

4. Scientific writeup — Level 3 (AI Collaboration) Your interpretation of the results. Where your verification evidence changed a conclusion or caught an error, that must be reflected here. The argument, what your results mean, why they are or are not what you expected, what caveats apply must remain yours.

All AI-generated content must be cited. Your self-assessment for this project should address where your trust map turned out to be wrong and what that tells you about calibrating this kind of judgment in future work.

Self-assessments — Level 1 (No AI) Self-assessments are the one place where AI use is not permitted. The honest reflection is the work. Using AI to write or refine your self-assessment defeats its entire purpose.

Project Grading

Each project asks you to write sections of a scientific publication, produce relevant figures and visualizations, and explore discussion questions designed to extend your thinking beyond the immediate analysis. The emphasis throughout is on growth, not performance.

Feedback, not scores. After each project you will receive detailed written feedback tied to the learning objectives for that report. You will also receive an indicative grade, but treat that number as a rough signal, not the point. The feedback is the point.

Growth model. No project grade is final until the end of the semester. Grades improve as you demonstrate that you have incorporated feedback from earlier reports into later ones. A weak Project 1 followed by clear improvement carries more weight than a strong Project 1 followed by stagnation.

Self-assessment. Each submission includes a self-assessment (AIAS Level 1 — see above). This is not graded on content. It is graded on depth and honesty. The questions are simple: where did you meet the learning objectives? Where do you still have gaps? What would you do differently? These reflections are the primary record of your learning across the semester. My hope is that this structure frees you to engage with AI tools transparently, focus on the material, and reflect honestly on your own growth without the grade getting in the way.

Final weights. Projects account for 80% of your final grade; lab participation accounts for 20%.

Course Schedule

Day Date Week Class Topic Project
Wed 9/2 1 Lecture Introduction  
Fri 9/4 1 Lab Lab 01 — Setup  
Mon 9/7   NO CLASS Labor Day  
Wed 9/9 2 Lecture Genomics, Genes, and Genomes
Next Generation Sequencing		 P1 assigned
Fri 9/11 2 Lab Lab 02 — Workflow Basics  
Mon 9/14 3 Lab Lab 03 — Nextflow Tooling  
Wed 9/16 3 Lecture Sequence Analysis Fundamentals  
Fri 9/18 3 Lab Lab 04 — Multi-Sample Pipelines  
Mon 9/21 4 Lecture Genomic Variation and SNP Analysis  
Wed 9/23 4 Lecture Long Read Sequencing  
Fri 9/25 4 Lab Lab 05 — Typed Channel Operators  
Mon 9/28 5 Lecture Sequence Analysis — RNA-Seq 1  
Wed 9/30 5 Lecture Sequence Analysis — RNA-Seq 2  
Fri 10/2 5 Lab Lab 06 — Containers (Docker)  
Mon 10/5 6 Lab Lab 07 — QC Pipeline with Singularity  
Wed 10/7 6 Lecture Biological Databases
Gene Sets and Enrichment		  
Fri 10/9 6 Lecture P1 Check-In and Review P1 due — P2 assigned
Mon 10/12   NO CLASS Indigenous People’s Day  
Tue 10/13 7 Lecture Genome Editing — CRISPR-Cas9
(Monday schedule substitute)		  
Wed 10/14 7 Lecture Sequence Analysis — ChIP-Seq  
Fri 10/16 7 Lab Lab 11 — RNAseq and DESeq2  
Mon 10/19 8 Lecture Sequence Analysis — ATAC-Seq  
Wed 10/21 8 Lecture P2 Check-In  
Fri 10/23 8 Lab Lab 09 — CRISPR Guide Design  
Mon 10/26 9 Lecture Microbiome: 16S and Metagenomics  
Wed 10/28 9 Lecture Metabolomics  
Fri 10/30 9 Lab Lab 12 — Differential Peak Analysis (ATACseq) P2 due — P3 assigned
Mon 11/2 10 Lecture Single Cell Analysis Part 1  
Wed 11/4 10 Lecture Single Cell Analysis Part 2  
Fri 11/6 10 Lab Lab 08 — Snakemake  
Mon 11/9 11 Lecture Single Cell Analysis Part 3  
Wed 11/11 11 Lecture Spatial Transcriptomics  
Fri 11/13 11 Lab Lab 10 — Genome Browsers  
Mon 11/16 12 Lecture P3 Check-In  
Wed 11/18 12 Lecture Single Cell Analysis Part 4 / Extended Topics  
Fri 11/20 12 Lab Lab 13 — Single Cell Setup P3 due — Final assigned
Mon 11/23 13 Lab Lab 14 — Single Cell QC  
  11/25   NO CLASS Thanksgiving Recess  
  11/28   NO CLASS Thanksgiving Recess  
Mon 11/30 14 Lab Lab 15 — Single Cell Preprocessing  
Wed 12/2 14 Lab Final Project Work Session  
Fri 12/4 14 Lab Lab 16 — Single Cell Pseudobulk  
Mon 12/7 15 Lab Single Cell Integration  
Wed 12/9 15 Lab Feedback  
  12/14     Final Exams Begin Final Project Due

A note on AI use in this syllabus

This syllabus was developed in keeping with the same principles outlined above. All content was initially drafted by me; Claude (Anthropic) was subsequently used to refine language, improve clarity, and suggest structural edits. Every element was reviewed, revised, and approved by me before inclusion.

The AI Assessment Scale sections were drafted with AI assistance and reviewed against the primary literature. Citations and framework descriptions reflect my own reading of that literature and should be verified against the original sources.

This syllabus was produced at approximately Level 3 (AI Collaboration): AI assisted in developing and refining the content while my judgment directed and evaluated the output throughout. Nearly all of the content was manually drafted first before integrating suggestions and edits from a LLM.

Unless otherwise specified, all other course materials including labs, projects, and supporting documents were produced at approximately Level 2–3 (AI Planning to AI Collaboration). I initially generated all drafts by myself without the use of AI and AI was used only for structure, organization, formatting, and further brainstorming. All of the technical content, biological framing, and pedagogical decisions remain my own.

In short, this document is my own work, produced with AI as a collaborator and not as a substitute for my judgment.