DNA Sequence Analysis Quiz

Computational Genome Analysis: Lecture 4

• A DNA sequence is presented, alongside questions relating to its analysis.
• The first question is about suitable statistical methods for describing the sequence.
• The second question asks what organism the sequence originated from.
• The third question examines if sequence parameters differ from bulk DNA parameters in the same organism.
• The fourth question explores the sequence type (e.g., protein coding, centromere, telomere, transposable element, control sequence).
• The lecture then focuses on analyzing short DNA strings (words) using k-tuple/k-mer analysis.

k=1 Analysis (Base Composition)

• For a DNA duplex, the number of As equals the number of Ts, and Gs equals Cs.
• These relationships hold true for the same strand.
• This concept is crucial for duplex DNA analysis, but not applicable to single strands.
• Base composition is a descriptive statistic widely used since the early days of molecular biology.

Biological Words & GC Content

• If a genome is GC-rich, the melting point will be higher than in an AT-rich genome due to stronger GC bonds.
• This difference in bond strength affects denaturation (strand separation).
• Organisms with high GC-rich content in their genome often inhabit hot springs.

Base Compositions of Various Organisms (Table)

• The G+C content varies among different organisms.
• Variations in GC content is due to factors such as selection, mutational biases, and biased recombination during DNA repair.

GC Skew

• GC skew is a useful way to describe the G-C balance in bacterial genomes.
• GC skew is useful in identifying replication origin and termini positions in prokaryotic sequences.
• GC skew is calculated for small windows along a genome.
• Third codon position GC skew (GCS) using windows of 300Kb and sliding 10Kb is commonly use in E. coli and B. subtilis to determine location of replication origins & termini.

Isochores (GC-rich Regions)

• GC content is not uniformly present in genomes.
• GC-rich regions called isochores are crucial, as they contain many protein-coding genes.
• About 50% of human genomes are GC rich.
• Isochores are large regions of DNA (>300kb) with homogenous G and C content.

Probabilistic Models

• A probabilistic model is useful in analyzing whether a DNA pattern occurs more frequently than expected by chance.
• This analysis can help determine if a pattern carries a biological significance.
• Probabilistic models simulate DNA sequences by defining probabilities for each base (A, C, G, T).

Expected Value & Variance

• The expected value (mean) and variance are key parameters in describing a distribution of a random variable.
• The expected number of times a specific nucleotide (letter) appears in a particular DNA sequence (e.g., number of A's) can be estimated.

k=2 Analysis (Dinucleotide Frequencies)

• Dinucleotides (e.g., AA, AC, AG, AT, …) are frequent in genomic analysis.
• The sum of the dinucleotide frequencies equals 1.
• Organism's "genomic signature" is the set of di-nucleotide frequencies which are useful for identifying horizontally transferred regions.
• A chi-squared test can be used if the observed number of dinucleotides differ from the theoretical expected number.

k=3 Analysis (Codon Frequencies)

• There are 61 codons in a standard genetic code.
• The usage of synonymous codons varies for amino acid types.
• This leads to "bias" in the codon frequencies in highly expressed genes.
• Statistical descriptions to the frequency variation in codon frequencies is important.
• Codon Adaptation Index (CAI) is a widely used measure for determining the gene expression level in DNA sequences.

k-tuples (k>3)

• Larger k-tuples (k ≥ 4) have important in genomics, and can be useful in identifying restriction sites, structural variations, and even determining if a sequence is coding or noncoding.
• For example, the Chi sequence (GCTGGTGG) is over-represented in bacterial genomes.

Summary and Applications

• GC content, GC skew, and k-mer distributions are frequently used in locating functional regions of DNA.
• Genomic analysis studies help determine if a sequence is coding or non-coding.
• Various parametric and non-parametric methods to study k-tuples are used.
• Applying these methods to predict gene expression, and identifying highly expressed genes, determining the functionality of a sequence are examples of applications.
• The methods can predict regions that have been taken up from other regions/organisms, regions involved in recombination and transcription, etc

Mystery of the Chilean Blob

• A 13-tonne blob washed ashore in Chile posed a biological mystery
• Hypotheses included various organisms, including giant squid.
• Ultimately, DNA analysis confirmed the blob was sperm whale blubber