Bioinformatics Test 3 (Wk 4-10)

Linux, Workflows & HPC. Genome Assembly and Assembly Quality. Genome Annotation. Sequencing and Variants. Variant Consequences. Single-Cell Transcriptomics. Protein Bioinformatics.

What is Linux?

An operating system widely used for bioinformatics servers and computing.

Why is Linux commonly used in bioinformatics?

It is stable, scalable, open-source and highly scriptable.

What is the command line?

A text-based interface for directly controlling the operating system.

What does ls do?

Lists files and directories

What does ls -a do?

Lists all files, including hidden files

What does cd do?

Changes the current directory

What does pwd do?

Shows the current working directory

What does grep do?

Searches and filters text

What does ssh do?

Connects to a remote computer

What does * mean as a Linux wildcard?

Matches multiple characters/files

What does ? mean as a Linux wildcard?

Matches a single character

What does | do in Linux?

Pipes the output of one command into another

What do wildcards and pipes enable?

Efficient processing across many files

What is a bioinformatics workflow/pipeline?

A sequence of analysis steps used to process biological data

Why do workflows improve reproducibility?

Steps and settings can be consistently repeated

What are three benefits of modern workflows?

Reproducibility, efficiency and scalability

Name three workflow management systems

Galaxy, Nextflow and Snakemake

What does HPC stand for?

High-Performance Computing

What is HPC used for?

Analysing very large datasets using many CPUs and large memory

What are the main components of an HPC cluster?

Login nodes, compute nodes, scheduler and shared storage

Where should HPC jobs actually run?

Compute nodes, after submission through the scheduler

What is SLURM?

A scheduler used to manage HPC jobs

What is Bash used for?

Command-line scripting and automation

What is Python commonly used for?

Automation, file processing and workflow integration

What is R mainly used for?

Statistical analysis and data visualisation

What does Conda manage?

Software environments and package dependencies

Why are containers useful?

They package software and dependencies to improve reproducibility

What is genome assembly?

Computationally reconstructing a genome from sequencing reads

What is the general assembly progression?

Reads → contigs → scaffolds → chromosomes

What is a sequencing read?

An individual sequence produced by a sequencing experiment

What is a contig?

A contiguous sequence assembled from overlapping reads

What is a scaffold?

Ordered/oriented contigs that may contain gaps

What is a chromosome-level assembly?

An assembly with chromosome-scale sequence continuity

What is de novo assembly?

Assembly performed without using a reference genome

What is resequencing?

Sequencing and analysing a genome against an existing reference

What is shotgun sequencing?

Randomly fragmenting the genome and computationally assembling it

Why are repetitive regions difficult to assemble?

Short reads may not span repeats, causing ambiguous placement

Why are long reads useful for genome assembly?

They can span repetitive regions that challenge short reads

What are tandem repeats?

Repeated sequences occurring next to one another

Give examples of tandem repeats

Satellites, minisatellites and microsatellites

What are interspersed repeats?

Repeated elements distributed throughout the genome

Give examples of interspersed repeats

LINEs, SINEs, LTR elements and DNA transposons

What does N50 measure?

The sequence length where 50% of the assembly is in sequences that long or longer

What does N50 tell you about an assembly?

Its contiguity

What does L50 measure?

The number of contigs/scaffolds needed to represent 50% of the assembly

What is the easiest way to remember N50 vs L50?

N50 = length; L50 = number

What does NG50 use instead of assembly size?

The known/expected genome size

What does BUSCO assess?

Presence and completeness of expected single-copy genes

What does a low BUSCO score potentially indicate?

Missing or fragmented expected genes/regions

What does N50 measure compared with BUSCO?

N50 = contiguity; BUSCO = completeness

What is genome annotation?

Identifying genomic features and assigning biological meaning to DNA

What is structural annotation?

Identifying where genes and other genomic features occur

What is functional annotation?

Assigning biological information such as function and pathways

What is the easiest way to remember structural vs functional annotation?

Structural = WHERE; functional = WHAT it does

Why is eukaryotic gene annotation difficult?

Introns, exons, UTRs, ncRNAs, non-coding DNA and complex transcription

Why are prokaryotic genomes generally easier to annotate?

Their gene structures are generally simpler and lack extensive intron/exon organisation

What are intrinsic (ab initio) gene predictions based on?

Information within the DNA sequence itself

What can intrinsic prediction use?

Codon usage, splice sites, branch sites, polyA signals, TSS and promoters

What are extrinsic/evidence-based predictions based on?

External DNA, RNA, protein similarity and synteny

What is integrated gene prediction?

Combining intrinsic and extrinsic evidence

What is AUGUSTUS used for?

Eukaryotic ab initio gene prediction

What is MAKER2 used for?

Integrating multiple evidence sources into genome annotations

What does RepeatMasker identify?

Repetitive elements

What is GFF3 used for?

Storing genome annotation features and their coordinates

Why shouldn't one gene prediction tool be trusted alone?

Predictions can disagree and contain exon, UTR or gene-calling errors

What does FASTA contain?

Sequence information with a header

What does FASTQ contain that FASTA does not?

Quality scores for each base

What are the four lines of a FASTQ record?

Identifier, sequence, + separator, quality scores

Which FASTQ line contains quality scores?

The fourth line

What does a Phred score represent?

The probability that a base call is incorrect

What does Phred 30 mean?

A 1 in 1000 chance the base call is incorrect

What are the two main ways to improve FASTQ quality?

Trimming and filtering

What does FastQC do?

Assesses sequencing quality without modifying the data

What is the general variant-calling workflow?

FASTQ → QC → mapping → SAM/BAM → variant calling → VCF

What does SAM contain?

Alignment coordinates and sequences of reads mapped to a reference

What is BAM?

The binary form of SAM

What is VCF used for?

Storing called variants and related information

What does QUAL mean in a VCF?

Phred-scaled quality score for the variant call

What are the main VCF fields?

CHROM, POS, ID, REF, ALT, QUAL, FILTER and INFO

What is an SNV?

A single nucleotide variant

What is an indel?

An insertion or deletion

What is a CNV?

A copy-number variant

What is a structural variant?

A larger-scale genomic variation such as a rearrangement

Why is high coverage useful in variant calling?

It increases confidence in distinguishing true variants from errors

What is strand bias?

Variant evidence being disproportionately supported by one sequencing strand

What forward/reverse support is preferable for a real variant?

Evidence from both forward and reverse strands

What is a synonymous variant?

A nucleotide change that does not alter the amino acid

What is a missense variant?

A nucleotide change that alters an amino acid

What is a nonsense variant?

A variant that introduces a stop codon

What do HGVS prefixes g., c. and p. mean?

g. = genomic; c. = coding DNA; p. = protein

What is dbSNP mainly used for?

Small-scale genetic variants

What is dbVar mainly used for?

Structural variants

What is ClinVar mainly used for?

Clinical significance of genetic variants

What is VEP used for?

Predicting variant consequences

What is bulk transcriptomics?

Measuring gene expression from an entire tissue or mixed cell population

What is the main problem with bulk transcriptomics?

Different cell types contribute mixed expression signals

What is single-cell transcriptomics?

Measuring gene expression on a per-cell basis

Why is single-cell transcriptomics useful?

It distinguishes expression patterns between individual cell populations

What evidence can be combined to infer protein function?

Sequence, domains, topology, structure, proteomics and pathways

What is a protein domain?

A larger structural/functional unit that can often fold independently