🐛Biostatistics Unit 14 – Genomic Data Analysis and Bioinformatics

Key Concepts and Terminology

• Genomics studies the structure, function, evolution, and mapping of genomes
• Bioinformatics combines biology, computer science, and statistics to analyze and interpret biological data
• Nucleotides (adenine, guanine, cytosine, and thymine) are the building blocks of DNA
• Genes are segments of DNA that encode instructions for making proteins or functional RNA molecules
• Transcriptomics examines the RNA transcripts produced by the genome at a given time
• Proteomics studies the structure, function, and interactions of proteins
• Metabolomics investigates the small molecule metabolites in cells, tissues, or organisms
• Epigenetics explores heritable changes in gene expression that do not involve alterations to the DNA sequence

Biological Data Types and Formats

• DNA sequencing determines the order of nucleotides in a DNA molecule
  • Sanger sequencing is a traditional method that uses dideoxynucleotides to terminate DNA synthesis
  • Next-generation sequencing (NGS) technologies enable high-throughput, parallel sequencing of millions of DNA fragments
• FASTA format is a text-based format for representing nucleotide or amino acid sequences
  • Begins with a ">" symbol followed by a sequence identifier
  • Sequence data follows on subsequent lines
• FASTQ format stores both biological sequence and its corresponding quality scores
  • Quality scores indicate the reliability of each base call
• SAM (Sequence Alignment/Map) format is used to represent aligned sequence reads
  • Contains information about read alignment positions, quality scores, and flags
• BAM (Binary Alignment/Map) is the binary, compressed version of the SAM format
• VCF (Variant Call Format) is a text file format for storing genetic variant information

Data Preprocessing and Quality Control

• Raw sequencing data must be preprocessed and quality-controlled before analysis
• Adapter trimming removes adapter sequences introduced during library preparation
• Quality filtering removes low-quality reads or trims low-quality bases
  • Phred quality score measures the probability of an incorrect base call
• Read deduplication identifies and removes PCR duplicates
• Contaminant screening detects and filters out contaminating sequences (non-target organisms)
• Quality control metrics assess the overall quality of the sequencing data
  • Per-base quality scores, GC content, sequence duplication levels, and overrepresented sequences
• FastQC is a widely used tool for generating quality control reports

Sequence Alignment and Assembly

• Sequence alignment compares and aligns two or more sequences to identify similarities and differences
  • Pairwise alignment compares two sequences (global or local alignment)
  • Multiple sequence alignment aligns three or more sequences
• Sequence assembly reconstructs the original DNA sequence from fragmented sequencing reads
  • De novo assembly builds the sequence without a reference genome
  • Reference-guided assembly aligns reads to a known reference genome
• Burrows-Wheeler Transform (BWT) is a data compression algorithm used in many alignment tools
• Alignment tools (BWA, Bowtie2, HISAT2) efficiently align sequencing reads to a reference genome
• Genome assemblers (SPAdes, Velvet, SOAPdenovo) reconstruct the original sequence from overlapping reads

Genomic Variant Detection

• Variant calling identifies differences between the sequenced genome and a reference genome
• Single nucleotide polymorphisms (SNPs) are single base pair changes
• Insertions and deletions (indels) are the addition or removal of one or more nucleotides
• Structural variations (SVs) include large-scale changes (copy number variations, translocations, inversions)
• Variant callers (GATK, SAMtools, FreeBayes) detect and genotype variants from aligned sequencing data
• Variant annotation adds biological context to identified variants (gene names, functional consequences)
• Variant filtration removes low-quality or likely false-positive variants based on predefined criteria

Statistical Methods in Genomics

• Hypothesis testing assesses the statistical significance of observed differences
  • T-tests compare means between two groups
  • ANOVA tests for differences among three or more groups
• Multiple testing correction adjusts p-values to control for false positives when conducting numerous tests
  • Bonferroni correction divides the significance threshold by the number of tests
  • False Discovery Rate (FDR) controls the expected proportion of false positives among significant results
• Differential expression analysis identifies genes with significant expression changes between conditions
  • DESeq2 and edgeR are popular R packages for differential expression analysis of RNA-seq data
• Enrichment analysis determines if a set of genes is overrepresented in a particular biological pathway or function
• Clustering algorithms (hierarchical, k-means) group similar samples or genes based on their expression profiles

Machine Learning in Bioinformatics

• Supervised learning trains models on labeled data to make predictions or classifications
  • Classification algorithms (logistic regression, support vector machines, random forests) predict discrete classes
  • Regression algorithms (linear regression, elastic net) predict continuous values
• Unsupervised learning discovers patterns or structures in unlabeled data
  • Dimensionality reduction techniques (PCA, t-SNE) visualize high-dimensional data in lower-dimensional space
  • Clustering algorithms (k-means, hierarchical clustering) group similar samples or features
• Feature selection identifies informative variables for model training
  • Univariate feature selection ranks features based on their individual relationship with the target variable
  • Regularization methods (LASSO, Ridge) perform feature selection during model training
• Cross-validation assesses model performance and prevents overfitting
  • Data is partitioned into training and validation sets multiple times to evaluate model generalization

Visualization and Interpretation of Results

• Heatmaps display patterns in gene expression or other genomic data across samples
• Volcano plots visualize differentially expressed genes based on fold change and statistical significance
• Manhattan plots display the genomic position and significance of genetic variants
• Pathway diagrams illustrate the relationships and interactions among genes or proteins in biological pathways
• Gene Ontology (GO) terms describe gene functions and enable functional annotation of gene sets
• Kyoto Encyclopedia of Genes and Genomes (KEGG) is a database of biological pathways and molecular interactions
• Interpretation of results should consider biological context, study design, and potential limitations
  • Integrate findings with existing knowledge and literature
  • Assess the reproducibility and generalizability of the results

Ethical Considerations and Data Privacy

• Informed consent ensures that participants understand the risks and benefits of sharing their genomic data
• Data anonymization removes personally identifiable information to protect participant privacy
• Data access control restricts who can view and use sensitive genomic data
  • Tiered access systems grant different levels of access based on user credentials and data sensitivity
• Genetic discrimination occurs when individuals are treated differently based on their genetic information
  • Genetic Information Nondiscrimination Act (GINA) prohibits discrimination in health insurance and employment
• Incidental findings are unexpected discoveries with potential health implications for participants
  • Researchers should have a plan for handling and communicating incidental findings
• Responsible data sharing enables scientific collaboration while protecting participant privacy
  • Data repositories (dbGaP, EGA) facilitate controlled access to genomic datasets