Resources & Practice

Learning bioinformatics becomes much easier when you follow a structure: study one concept, run one small analysis, explain the output, and keep a short written record. This page collects practical study routes and reporting guidance so your progress feels deliberate rather than scattered.

Suggested study roadmaps

Roadmap 1

General NGS learner

Best if you want a strong overview of common sequencing workflows.

Getting started -> Data formats -> QC -> Alignment -> Variant calling -> Pipelines

Roadmap 2

Transcriptomics learner

Best if your work is centered on expression, cell states, or treatment response.

QC -> RNA-seq -> Single-cell RNA-seq -> Practice lab

Roadmap 3

Assembly & workflow learner

Best if you care about reproducible pipelines, assembly, and project organization.

Assembly -> Pipelines -> Tools reference -> Practice lab

Practice exercises that build intuition

QC translation: run FastQC or MultiQC and explain five panels in plain English.
Trimming tradeoff: compare before/after length distributions and explain why aggressive trimming may lose useful reads.
Alignment summary: record mapping rate, duplicate rate, and insert size; then explain what each metric means biologically.
Variant filtering: change DP/GQ/VAF thresholds and justify why each threshold exists.
RNA-seq interpretation: write a short report that combines PCA, volcano plot, and three biologically relevant genes.
Single-cell QC: vary cell filtering thresholds and ask how many cells are lost versus how much noise is removed.

A lightweight reporting template

Question: what biological problem are you trying to answer?
Data: source, accession, platform, and sample manifest
References: genome build, annotation, database versions
Workflow: tools, versions, and key parameters
QC decisions: which plots mattered and what thresholds were used
Main results: 2-4 core plots and summary tables
Limitations: what remains uncertain or could bias interpretation

How to turn a tutorial into real skill

Repeat on a second dataset

If your understanding is real, you should still recognize the same metrics and pitfalls on a different dataset.

Explain one figure aloud

Teaching a plot to someone else is one of the fastest ways to discover whether you actually understand it.

Write assumptions down

If you changed a threshold or chose a tool, record why. Reproducible reasoning matters as much as reproducible code.

Command-line mini cheat sheet

# FASTQ
zcat reads.fastq.gz | head
expr $(zcat reads.fastq.gz | wc -l) / 4

# BAM
samtools view -H sample.bam | head
samtools flagstat sample.bam

# VCF
bcftools view -h sample.vcf.gz | head
bcftools view -H sample.vcf.gz | head

# Counts / matrices
head counts.tsv
cut -f1-5 counts.tsv | column -t | head

Where to continue

Practice Lab for browser-based hands-on exercises
Tools Reference for a curated catalog, including FastQLab and JapalitySplice in proper workflow context
Pipelines for reproducibility and workflow design

A good self-check question

Can you explain why a specific sample passed or failed QC, what file carried the result forward, and which downstream conclusion would change if that decision were different? If yes, you are learning the right way.