Lecture 6

• description of gene function

• restricted vocabulary

• hierachical organized

• directect acylic graph

• 3 sub-ontologies

• evidence codes

we can identify the taxonomy

database of molecular function represented in terms of functional othologs

• contains two uniques aspects, KO assignments and KEGG mapping

  
  

  1. KO assignment: Molecular function are stored in the KO (KEGG orthology) databse containing orhtologs of experimentally characterized genes/proteins

  • Genome annotation in KEGG is to assign identifiers (or K numbers) to individual genes in the genome, rather than giving text description of functions

1. KEGG mapping: cellular and organism-level function are stored in the Pathway,Brite and Module databse in terms of the molecular network, which are all created as networks of K number nodes

   • The KO assignment procedure converts a gene set in the genome to a K number set and leads to automatic recosntrcution of KEGG pathways and other networks by the process call KEGG mapping, enabling interpretation of high level functions

• a curated resource devoted to proteins and their function

• comprising many different databases, of which the KEGG Orthology database is the most basal annotation layer

• the KO assigns orthologs to experimentally characterized genes the same KO number, i.e. assigns the as functional equivalents

• the pathway database integrates gene into higher-order functional units

• a pathway is a network if KO numbers

• assigning a KO number to a protein resembles

  • a functional annotation transfer

  • an integration into corresponding pathway

1. significant sequence similarity (sequence conservation)

2. orthology relationships - descendants of the same gene in the last common ancetsral species

3. conserved genomic position (positional homologs/orthologs)

4. similar/identical domain architectures

5. similar/identical 3D strcutures

6. agreeing (conserved) expression pattern

7. Interaction partners present / conservation of interaction networks

• homology is only about evolutionary relationships

• sequence similarity is not involved in the concept of homolgy. It measures the extent to which two sequences agree

• Homolgy is - in the first - approximation - a Yes/no state

• similarity can be measured in the percent

• homology is not about function

1. Sequence space is tremendous

2. sequence are made up from infinite alphabet

3. from 1 and 2 follows the sequence are not related are no similar than it is expected by chance. Extent of similarity by chance follows a random distribution

4. From 1 to 3 does not follow that sequences that are homologus must share a significant sequence similarity. Remember, sequence similarity decays with time, the relatedness of sequences doesnt change

• gene duplication

• the reciprocal best blast hit approach

• instead of comparing proteins pair-wise, we want to get an overview about as many proteins as possible in the clade of interest

  > reduce the impact of chance and noise

  • which species harbour an ortholog to your gene of interest

  • how does the ortholog “look like”

• False positive

  • contamination

  • limited specificity

• False negatives

  • data quality

  • limited senstivity

• misleading interpretation of true positives

  • lineages specific change of function

• size of the matrix - how many taxa do we need