Gene Ontology

inititiative to unify the representation of gene and gene product attributes across all species

provides controlled vocabulary for gene function

includes hierarchical description of relationships among the terms (organization of genes)

Genome Annotation

process of identifying locations of genes and coding regions and determining what those genes do

for example, once the genome is sequenced, it needs to be annotated to make sense of it

Systems biology

field that studies the interactions and relationships between various parts of a biological system and to integrate the information to better understand how the systems functions

Example of hierarchical GO terms

1. function

2. metabolism

3. TCA cycle enzymes

how many genes in a human?

35K

Genome scales

microbes

C. elegans

Drosophila

Arabidopsis

Fish

(give genome size and % of disease genes shared)

microbes ~6K; 20%

C. elegans ~18.5K; 60%

Drosophila ~13.5K; 58%

Arabidopsis ~25.5K; 48%

Fish 35K; >75%

Phenome

set of all phenotypes expressed by cell, tissue, organ, organism

only in the model organisms

Emergent properties

example

networks exhibit emergent properties--# of simple entities (agents, genes) operate in an environment, forming more complex behaviors as a collective

example Robustness

network is made up of

nodes and edges

Random network

bell curve

every node has about the same # of links

no highly connected nodes

Scale free network

this architecture can happen across many different scales--don't needs a certain number of nodes to give it structure

Power Law Distribution

few hubs with large number of links

many nodes with small # of links

network structure tells you about

network function

are random networks or scale free networks more robust

(how robust is it in face of random failures at nodes)

give exception to this rule

scale free networks more robust

network will not collapse in face of most random failure

UNLESS a hub goes out

biological networks and gene networks are

scale free

this is why we can live with some mutations

the network structures serves as a buffer system

holy grail of systems biology

collect data,

integrate it,

model it;

use info to predict how system reacts to untested condition

give four examples of edges in biology

enzymatic reactions

protein: protein

protein: DNA

miRNA; RNA

three different types of nodes in biology

genes, proteins, metabolites

four classifications of different edges

metabolic

regulatory

co-expression

protein interaction

which two classifications of edges can have direction?

metabolic (e.g. enzymatic pathways that require ATP are usually only in one direction)

regulatory

which edge classification can be weighted

co-expression

(transcriptome date; weight based on correlation of gene expression)

network definitions:

Degree

number of edges connected to a node

in degree

# of incoming edges

out degree

number of outgoing edges

connectivity index

tells you what

who are the important players

more connections, higher biological significance value

polycistronic RNA

multiple genes transcribed as one transcript

why is transcriptional control most important level of regulation in bacteria?

bacteria have no nucleus

transcription and translation can occur simultaneously

process of finding significant network motifs

intentify all TFs and targets

generate random network ensemble (reference network) with same number of nodes and connections

count how many times motif appears in real vs random network

compute statistical significance for each motif

statistical significance for each motif

probability of a network motif appearing in random network as frequently as in real network

sub-graph

motif

p value

p-value is the probability of obtaining a test statistic at least as extreme as the one that was actually observed

The lower the p-value, the less likely the result, so the more "significant" the result, in the sense of statistical significance

z score

The number of standard deviations from the mean is 

(aw data value minus the mean and then this result divided by the standard deviation)

the higher the better

feed forward loop function

also called one-two punch

noise filter

(filters out transient signals / sporatic changes in environment)

it is a persistence detector

requires SUSTAINED input

sign sensitive (positive or negative) delay

network motifs

feed forward loop

single input module

dense overlapping regulon

dense overlapping regulon

lots of TFs which activate a bunch of genes

edges overlap

you have and / or functions

Regulon

collection of genes or operons under regulation by the same regulatory protein

E. Coli transcription network

shallow network

with Dense overlapping regulon modules

shallow network

one layer

TFs targeting operons

(not layers and layers of TFs regulating each other)

stimulon

group of regulons responding to a single stimulus

why are network motifs important in yeast and other eukaryotes

genes in similar pathways not organized into operons

need to be coordinately expressed

monocistronic

each gene has its own RNA transcript

interactome

all the interactions between biological entities in cells and organisms considered as a whole

how to create an interactome map?

yeast two hybride (binary interactions)

mass spectrometry (co-IP, ChIP chip)

limitation of yeast two hybrid system

limited to looking at binary interactions

get false positives (interact in vitro but not in vivo)

yeast two hybrid screens

performed in yeast and what other model

cloning and plasmid preparation performed in E. coli

(hubs)

positive correlation between connectivity and

lethality

more highly connected hubs in network are more likely to be essential

2 different kinds of hubs

date hubs

party hubs

Party hubs

function inside each module

same time and space

high correlation of expression with connected genes

Date hubs

connect distinct modules

different time and/or space

low correlation of expression with connected genes

Pearson correlation coefficient

is a measure of the correlation (linear dependence) between two variables X and Y, giving a value between +1 and −1 inclusive.

widely used as measure of the strength of linear dependence between two variables

correlated: value of 1

less correlated 0.5

higher pearson correlation coefficient

higher correlation of expression

Auxotrophic mutant

mutation that renders organism unable to synthesize an essential compound

replica plating

technique in which one or more secondary Petri plates containing different solid selective growth media are inoculated with the same colonies of microorganisms from a primary plate (or master dish), reproducing the original spatial pattern of colonies.

Galactose utilization pathway in yeast

galactose activates GAL3

GAL3 respressed GAL80

which relives GAL80's repression of GAL4

GAL3

galactose activates it

GAL3 represses GAL80

GAL80

repressed by GAL3

GAL80 represses GAL4

GAL4

TF that activates several metabolic genes

cytoscape

software which allows you to visualize network

view data as nodes and edges

Systems approach to regulation of GAL metabolic pathway

discovered what

regulatory network extends way beyond what was previously thought

uncovered new players in GAL universe

new links btw regulation of GAL metabolic pathway and cellular processes

Multinetwork

network where multiple types of edges link genes

how can you draw protein:DNA edges?

based on correlated expression of TFs and target genes

based on cis data (binding sites)

Geneways

text mining tool for gene interactions

interolog

uses homology between species to make prot:prot interaction predictions

weight of interaction edge btw genes is based on BLAST score of the orthologs

KEGG

database for metabolic reactions

biochemical pathways

ortholog

A gene in two or more species that has evolved from a common ancestor.

CCA1

transcription factor

regulates circadian clock in plants

if mutation of two different genes gives you the same phenotype, what can you do?

draw edge between 2 genes

molecular machines

subset of genes that work together to build embryo

fate map

technique that is used to show how a cell or tissue moves and what it will become during normal development.

representation of the developmental history of each cell in the body of an adult organism

lets you id genes for embryo development

genes in same phenocluster tend to be

closer together in the interaction network than expected by chance

Directed graph

ordered pairs of vertices

Robustness

keeping your composure in changing conditions

transcriptome / co-expression data

mRNA correlation

weighted

motifs in nework can give clues to what

biological processes in which they occur

sign-sensitive delay

FFL

a circuit that responds rapidly to step-like stimuli in one direction (e.g. ON to OFF), and at a delay to steps in the opposite direction (OFF to ON)

Single Input Module

(SIM)

single TF regulates several genes

all regulation are of the same sign

how are date and party hubs distinguished?

on the basis of their partner's expression profiles

which types of hubs are more central to network topology?

date hubs

which types of hubs have higher pearson correlation coefficient?

party hubs

what did the discovery of two types of hubs tell us about the organization of the yeast proteome? (Marc Vidal paper)

the results support a model of organized modularity for the yeast proteome

date hubs are connected to modular subnetworks

proof of principle

demonstration of the commercial potential of a discovery or invention

scoring phenotypes

way to compare phenotypes quantitatively

after certain gene mutation, first cell division DOESN"T happen: score 1

after certain gene mutation, first cell division DOES happen: score 0

Co-IP

an antibody raised to a bait protein binds the bait together with any other prey proteins that interact with it

the interacting proteins can be purified and analysed by western blotting or mass spec

TransFact DB

validated TFs and their targets

BIND or DIB database

protein protein interactions