Skip to content

UZH BIO390 - Learning Goals

This page indicates some of the learning goals, as emphasised by the different lecturers. Some points will have been discussed in different lectures; accordingly, exam questions may not refer to information of one specific presentation.

Learning goals have a general scope

Please be aware that some of the "Learning Goals" may reflect aspects not necessarily captured by the lectures in the current semester - The ones relevant for the current semester's exam are related to the given lectures. Also, updates may occur at any time.

Consider individual pages

Please consider individual course pages and documents linked from there. Those might be updated later on, so it is a good idea to check back before the end of the course.

Bioinformatics: Definition & Concepts

  • definition of "Bioinformatics" (cf. Anna Tramontano)
  • categories of informatics tools used in bioinformatics
  • hypothesis versus data driven science
  • areas of bioinformatics/bioinformaticians, (in contrast to "pure" modelling, statistics etc.)
  • 3 main categories of biological data, and example resources
  • definition of API
  • common sequence related file formats
  • hierarchies and relationships as 2 main principles of ontologies
  • areas of "not-bioinformatics", and why
  • bioinformatics tools (programming languages, libraries, online resources) and their specific use cases

Sequence Analysis

  • basics of DNA and Protein sequences
  • substitution matrices
  • BLAST
    • parameters
    • terms
    • scores

Statistical Bioinformatics & Machine Learning

  • usage of gene expression profiling
  • multiple testing correction
  • parameters for hierarchical clustering
  • statistical evidence for a change in the mean
  • dimensionality reduction
  • central limit theorem
  • hierarchical clustering
  • clustering coefficient
  • unsupervised machine learning tasks/components
  • ML model types

Bioinformatics tools & resources

  • types of bioinformatics resources
  • common programming/analysis languages/tools in bioinformatics resource pipeline
    • e.g. R, Perl, Python, JavaScript, regex, git (version control) ...
  • components of bioinformatics online resources
    • Databases, webserver gateway, APIs, frontends ...
  • database types / concepts: SQL vs. NoSQL (e.g. document-oriented databases, graph database)
  • metadata and data curation (biocuration, codification)
    • importance of classifications, ontologies
    • some ontologies and their use (NCIt, UBERON, ICD-O, GO, PO)

Some Q & A (thanks to the providers of these questions)

  • Progenetix use case: In comparative genomic hybridization, in the case of an high copy number segment of DNA, more tumor DNA will hybridize to the metaphase chromosomes just because of higher likelihood?
    • in essence, yes; it is a mix of higherlikelihood and therefore higher binding probability (it is actually hard for a given fragment to encounter the right place on the chromosome or array) w/ or w/o competition effects (latter when using normal reference DNA)
  • CURIEs: ...hierarchical coding systems where individual codes are represented as CURIEs - aren't they a type of URI rather then a "code"?
    • CURIEs are universal identifiers (URIs) consisting of a public and a local part
    • they are universal (like UUIDs) but unlike UUIDs (which can be anonymous) they are resolvable (i.e. the public part can be resolved to a URL where then the local part can be used to retrieve the resource)
    • the "hierarchical coding systems" usually don't use the CURIE internally but only the private part; but using the complete CURIE externally makes it unambiguous
  • Progenetix use case: In Progenetix one can either use the GA4GH-Beacon API to query (i.e., do information retrieval) OR use the pgxRpi API to load the data into an analysis environment (i.e., for eventual knowledge extraction)?
    • Not really. The bycon software stack implements database access / middleware / Beacon API instance. This (outward facing) API (compatible to the Beacon specification) can be accessed by various clients for data retrieval (e.g. the beaconplus-web or progenetix-web JavaScript front ends, manual http requests, Beacon aggregator services...).
    • The pgxRpi itself is an API for the R environment, i.e. another client accessing the Beacon API. So here one gets (DATA - Progenetix' Beacon API - pgxRpi API - R analysis environment.

Regulatory Genomics and Epigenomics

  • secondary/tertiary human genome structure
  • functional genome content
  • transcription factors & genome interaction
  • chemical genome modifications, their effectors and results
  • Chip-Seq
  • read mapping
  • peak calling
  • sequence compression algorithms

Metagenomics

  • concept of taxonomic diversity
  • concept microbial community dissimilarity
  • how are sequences used to derive an adopted species concept for prokaryotes
  • principle steps for 16S rRNA-based taxonomic composition analysis
  • essential steps of short sequencing read assembly into contigs and scaffolds
  • basic steps of metagenomic analysis: from raw reads to the reconstruction of genomic scaffolds

Proteomics

  • principles of proteome organization in the cell
  • key experimental and computational concepts for the collection and analysis of high confidence protein-protein interaction data
  • peptide fragmentation
  • target-decoy approach
  • protein quantification
  • MassSpec similarity queries

Biological Networks

  • Protein interaction and metabolic networks
  • databases and online resources for different types of pathway and interaction data
  • detection of protein complexes
  • graphs, nodes, edges, paths
  • geodesics, graph diameter
  • common types of degree distributions
  • adjacency matrix
  • shortest path matrix
  • assortative and disassortative graphs
  • community (module) detection
  • cliques
  • motifs, graph representations of metabolic networks

Text Mining

  • text mining pipelines & (current) common programs/applications
  • article/literature repositories (with focus on accessibility)
  • processing steps in text mining
    • stemming etc.
  • common problems in text mining
  • search engine precision metrics
  • benchmarking

Semantic web, RDF, Ontologies

  • semantic web
    • elements
    • benefits
  • components of ontologies
  • stack of standards in semantic web and their functions
  • RDF for modeling data
  • OWL/OBO for modeling a biomedical domain
  • querying knowledge graphs for answering biomedical questions

Clinical Bioinformatics & Personalized Medicine

  • genomic variants (types, numbers)
  • reference genome(s)
  • main bottlenecks of molecular diagnostics in the clinical setting
  • goals of many personalized health initiatives
  • currently favoured clinical NGS technology
  • clinical trial participation

Genomic data & privacy

  • reasons for needing many genomes
  • genomic privacy and re-identification (concepts)
  • principle of re-identification attacks over the Beacon protocol
  • long range familial searches
  • opinions about risk vs. opportunities
  • direct-to-consumer genetic testing -> what, how
  • technical and regulatory solutions against privacy breaches & data abuse
  • genotype-phenotype (G2P) (ab-)use