BMS5021 · Introduction to Bioinformatics
The Complete Study & Assessment Guide · S1 2027Introduction to Bioinformatics
BMS5021 Introduction to Bioinformatics is Monash University's postgraduate, Master-level introduction to computational biology, taught by Biomedicine Discovery Institute staff into the Master of Data Science and built for students with little or no prior biology. The Monash Handbook frames it around three topics delivered flipped-classroom style — asynchronous mini-lectures and eModules done before the weekly in-person workshops — moving from the biology of the gene and the genetic code, through the data formats and algorithms of sequence analysis (the command line, FASTQ quality, alignment, RNA-seq differential expression), to mass-spectrometry proteomics and multi-omics. Crucially, BMS5021 has no exam: it is assessed 100% by continuous coursework across the three topics. Topic 1 is three in-workshop summative MCQ quizzes worth 30% together; Topic 2 is a bioinformatics data-analysis exercise — an RNA-seq differential-expression pipeline on a colorectal-cancer dataset — submitted as a written report with an in-class presentation and a Topic-2 MCQ, worth 35%; Topic 3 is a quantitative-proteomics / mass-spectrometry data-analysis report worth 35%. That structure rewards steady, hands-on practice with real data and clear scientific writing rather than a last-minute cram, and the BMS5021 result feeds the Weighted Average Mark (WAM) that later units build on. No hurdle or competency threshold is stated in the unit materials — confirm the current assessment rules on Moodle and in the unit guide. BMS5021 is offered in both Semester 1 and Semester 2; the schedule shown in this guide follows the Semester-1 run.
What BMS5021 covers
BMS5021 has no exam — it is 100% coursework across three topics: Topic 1's three in-workshop MCQ quizzes (30% total), Topic 2's RNA-seq differential-expression report, in-class presentation and MCQ (35%), and Topic 3's quantitative-proteomics data-analysis report (35%). This eleven-chapter map follows the Semester-1 teaching schedule from DNA and the genetic code through sequence alignment, RNA-seq and mass-spectrometry proteomics, so you can see how each week's method feeds the data-analysis task it is marked on. Use it to keep the whole pipeline — raw reads to interpreted result — straight as you build each report.
How BMS5021 is assessed
| Component | Weight | Format |
|---|---|---|
| Topic 1 — Summative MCQ Quizzes | 30% | Three in-workshop online quizzes across Weeks 1-4 (~15% + 7.5% + 7.5%); the plagiarism & collusion statement must be completed before each quiz unlocks |
| Topic 2 — Data-Analysis Report + Presentation + MCQ | 35% | An RNA-seq differential-expression pipeline on a colorectal-cancer dataset: written report + in-class presentation + a Topic-2 MCQ (Weeks 5-8) |
| Topic 3 — Proteomics Report | 35% | A quantitative-proteomics / mass-spectrometry data-analysis report (Weeks 9-12) |
Reading an edgeR topTags row and the multiple-testing problem (Topic 2 data-analysis read)
- +1Read the sign and size of the log fold-change. logFC is measured on a log2 scale, relative to the reference (normal) group. A positive logFC = +7.42 means the gene is UP-regulated in the metastasized group; because it is log2, +7.42 corresponds to roughly 2^7.42 ≈ 171-fold higher expression than normal.
- +1Confirm it is a real call, not noise. The raw PValue = 0.00394 is below 0.05, and the adjusted FDR = 0.049 is still below 0.05, so the gene survives multiple-testing correction — it is a credible differentially-expressed gene, strongly up in metastasis.
- +1Count the expected false positives at the unadjusted threshold. If you test 20,000 genes and accept any gene with P <= 0.05 as a hit, you expect about 0.05 x 20,000 = 1,000 genes to pass by chance alone even if nothing were truly different. That is the multiple-testing problem.
- +1Say why FDR is the right column. The Benjamini-Hochberg FDR controls the expected PROPORTION of false discoveries among the genes you call significant, so it is calibrated for testing thousands of genes at once; the raw PValue is not. (Bonferroni FWER is the stricter, more conservative alternative that guards against ANY false positive.) You therefore rank and threshold on FDR, not the raw PValue.
Key terms
- Central dogma
- The flow of genetic information DNA -> (transcription) RNA -> (translation) protein, with replication of DNA and reverse transcription (RNA -> DNA) as additional routes. Gene expression is the multi-step reading of a gene to build its protein.
- FASTQ / Phred quality score
- FASTQ stores each read as four lines (identifier, sequence, +, per-base quality). The Phred score Q = -10 log10(P) encodes the probability P that a base call is wrong; Q30 = 1-in-1,000 error (99.9%) is the NGS quality benchmark.
- Sequence alignment (dynamic programming)
- Finding the arrangement of two sequences that maximises matching positions. Needleman-Wunsch gives the optimal GLOBAL alignment and Smith-Waterman the optimal LOCAL alignment, both by dynamic programming (initialise, fill, traceback) with a scoring matrix and gap penalties.
- RPKM / FPKM
- Reads (or Fragments) Per Kilobase of transcript per Million mapped reads: RPKM = (10^9 x C) / (N x L), where C = reads for the gene, N = total library reads, L = gene length in bp. Normalises a raw count for BOTH gene length and library depth.
- Differential expression (edgeR)
- Statistically testing, gene by gene, whether expression differs between conditions. edgeR outputs logFC (log2 fold-change vs the reference), logCPM, F, PValue and FDR; the null hypothesis is no difference for a given gene.
- Trypsin digestion / peptide-mass fingerprinting
- In MS proteomics, trypsin cleaves a protein C-terminally of arginine (R) and lysine (K), producing peptides whose masses are matched against an in-silico tryptic digest of a database to identify the protein by its peptide-mass fingerprint.
BMS5021 FAQ
Is BMS5021 hard?
It is broad rather than deep, and the challenge is that it mixes real biology with real computing. BMS5021 is built for students with little or no biology background, so the biology is taught from first principles; the harder adjustment for most is the command line, the R/edgeR and MSFragger workflows, and doing genuine data analysis on messy sequencing and mass-spec data. Because it is 100% coursework with no exam, success comes from keeping up with the flipped-classroom mini-lectures before each in-person workshop and practising the pipelines hands-on rather than cramming. Students who treat each week's method as something to run themselves, not just read, tend to find it manageable.
How is BMS5021 assessed?
There is no exam — BMS5021 is assessed 100% by continuous coursework across three topics. Topic 1 is three in-workshop summative MCQ quizzes worth 30% together (roughly 15% + 7.5% + 7.5% across Weeks 1-4). Topic 2 is worth 35%: a bioinformatics data-analysis exercise — an RNA-seq differential-expression pipeline on a colorectal-cancer dataset — submitted as a written report with an in-class presentation and a Topic-2 MCQ. Topic 3 is worth 35%: a quantitative-proteomics / mass-spectrometry data-analysis report. The intra-topic sub-weights come from Moodle rather than the Monash Handbook, so confirm the exact splits, dates and submission rules in the current unit guide on Moodle.
Does BMS5021 have a hurdle?
The Monash Handbook lists all three assessment components with a null hurdle field, and no threshold or competency hurdle is stated in the available unit materials — so on the current information there is no separate hurdle to clear beyond the overall pass mark. Because hurdle and pass rules can change between offerings, confirm the assessment rules, any threshold requirement and whether the quizzes are open- or closed-book on Moodle and in the unit guide before each task. Your BMS5021 result contributes credit points and feeds your WAM in the usual way.
Can AI help me with BMS5021?
Yes, as a study aid. Sia is an AI tutor trained on how BMS5021 is actually taught and assessed — it can explain the central dogma, walk a Needleman-Wunsch alignment or a Phred-score conversion step by step, unpack why you filter low-CPM genes before differential expression, or talk through interpreting a topTags row or a mass spectrum, and it checks your reasoning as you go. It does NOT do your graded quizzes, report or presentation for you, and Monash University academic-integrity rules apply. Note that AI-tool use in the written reports is governed per task by the educator, so follow the acknowledgement rules on Moodle.
What does it take to do well in the BMS5021 data-analysis reports?
Markers reward a clean, reproducible pipeline and interpretation you can justify, not just a result. For Topic 2, that means QC and trimming before alignment, sensible read-mapping and counting, correct normalisation (RPKM/FPKM or TMM), and differential expression reported on the adjusted FDR with the direction stated relative to the reference group and shown on a volcano plot. For Topic 3, it means defensible missing-value handling and quantification choices for the mass-spec data. Across both, structure the report clearly, label every figure, and explain WHY each step was done. Confirm the exact rubric and expectations for your cohort on Moodle.
How to prepare for the assessments
Treat BMS5021 as a set of hands-on pipelines to rehearse weekly, not a reading unit, and use the flipped-classroom rhythm as intended: do each week's asynchronous mini-lecture and eModule BEFORE the in-person workshop so you can spend the workshop actually running the analysis. For Topic 1, the three MCQ quizzes (30% across Weeks 1-4) reward knowing the biology cleanly — the central dogma, genome organisation (46 chromosomes, rRNA/tRNA/mRNA proportions), mutation classes, sequencing technologies and the proteome concept — so build a compact fact sheet and self-quiz. For Topic 2 (35%), practise the whole RNA-seq pipeline end to end on the colorectal-cancer dataset: QC and trim, align (Bowtie2/HISAT2), count with featureCounts, then edgeR (filter low CPM, TMM-normalise, fit, test) and read the topTags/volcano output critically. For Topic 3 (35%), rehearse the proteomics workflow — trypsin digestion, m/z and peptide-spectrum matching, quantification and missing-value handling — and be ready to justify every choice. Because everything is written up as a report or a presentation, keep a running note of WHY each step is done so your scientific reasoning is on the page. When a step won't click, ask Sia to explain that single step a different way and to set you fresh practice; it teaches the method and checks your reasoning, and it never substitutes for your own graded work. Confirm assessment dates, sub-weights and rules on Moodle.
Your AI Biology tutor for BMS5021
Stuck on a hard BMS5021 question? Sia is AskSia’s AI Biology tutor — ask any BMS5021 Introduction to Bioinformatics question and get a clear, step-by-step explanation grounded in how the course is actually taught and assessed. Read this whole study guide free, then take your hardest questions to Sia.
