14.04.2014 Views

Mathematical Modeling in Molecular Biology - Vision at IME-USP

Mathematical Modeling in Molecular Biology - Vision at IME-USP

Mathematical Modeling in Molecular Biology - Vision at IME-USP

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

<strong>M<strong>at</strong>hem<strong>at</strong>ical</strong> <strong>Model<strong>in</strong>g</strong> <strong>in</strong><br />

<strong>Molecular</strong> <strong>Biology</strong><br />

Junior Barrera<br />

DCC-<strong>IME</strong>-<strong>USP</strong>/BIOINFO-<strong>USP</strong>


Cell cycle,<br />

Tripanosoma-cruzi,<br />

Dictyostelium<br />

....<br />

CAGE<br />

Bioquímica-<strong>USP</strong><br />

Projects<br />

Cancer<br />

Virus<br />

NHGRI<br />

Hemocentro<br />

Agriculture<br />

Ribeirão-<strong>USP</strong><br />

VGDN<br />

ICB-<strong>USP</strong><br />

BIOINFO-<strong>USP</strong><br />

Agriculture<br />

ESALQ<br />

OMS<br />

ICB-<strong>USP</strong><br />

Malária<br />

ICB-<strong>USP</strong><br />

Cancer<br />

Ludwig


Layout<br />

• Economics perspective<br />

• Fundamentals of <strong>Molecular</strong> <strong>Biology</strong><br />

• Experimental techniques<br />

• D<strong>at</strong>a m<strong>in</strong><strong>in</strong>g environments<br />

• Some Biological problems<br />

• Formaliz<strong>in</strong>g Biological problems<br />

• A family of <strong>M<strong>at</strong>hem<strong>at</strong>ical</strong> problems<br />

• Other areas with similar problems


Economics Perspective<br />

• Atomic program (1930-1945)<br />

• Space program (1960-1975)<br />

• Genetics (1990- ...): <strong>in</strong>tern<strong>at</strong>ional<br />

<strong>in</strong>vestment, public and priv<strong>at</strong>e<br />

Brazilian <strong>in</strong>vestment is proportionally gre<strong>at</strong>er<br />

than the <strong>in</strong>vestment of some developed countries


Fundamentals of <strong>Molecular</strong><br />

<strong>Biology</strong>


Popul<strong>at</strong>ion<br />

organism<br />

Prote<strong>in</strong> structure<br />

and dynamics<br />

tissue<br />

DNA, Prote<strong>in</strong>,<br />

Gene Expression,<br />

Gene Networks<br />

molecule<br />

cell


Heredity - Mendel (1866)<br />

The phenotypes of an <strong>in</strong>dividual depends on genes of his<br />

parents.


Chromosome Theory - Morgan (1910)<br />

Genes were situ<strong>at</strong>ed <strong>in</strong> chromosomes


The molecular structure of chromosomes (W<strong>at</strong>son and Crick -<br />

1953)<br />

DNA structure: the double helix<br />

Four basis: aden<strong>in</strong>e(A), guan<strong>in</strong>e(G), thym<strong>in</strong>e(T), cytos<strong>in</strong>e(C)<br />

genes are sequences of nucleotides


DNA manipul<strong>at</strong>ion<br />

cut, replic<strong>at</strong>ion and decod<strong>in</strong>g


Genes control the metabolism<br />

Metabolism occurs by sequences of enzyme-c<strong>at</strong>alyzed<br />

reactions.<br />

Enzymes are specified by one or more genes


Gene expression


species modific<strong>at</strong>ion, diagnostics, drug production


Experimental techniques


Áreas de <strong>at</strong>uação


Microarrays


Image Analysis


Signal<br />

gene 1<br />

gene 2


D<strong>at</strong>a M<strong>in</strong><strong>in</strong>g Environments


Objected oriented d<strong>at</strong>abase<br />

P1<br />

P2<br />

Pn<br />

Pi : analytical and m<strong>in</strong><strong>in</strong>g procedures (kernel parallel)


Integr<strong>at</strong>ed Environment<br />

Genbank<br />

d<strong>at</strong>abase<br />

Another<br />

d<strong>at</strong>abases<br />

Access control<br />

module<br />

Cl<strong>in</strong>ical d<strong>at</strong>a<br />

Sequence module<br />

P1<br />

microarray module<br />

P1<br />

analitical<br />

oper<strong>at</strong>ional<br />

P2<br />

P3<br />

analitical<br />

oper<strong>at</strong>ional<br />

P2<br />

P3<br />

query<br />

Pn<br />

Pn<br />

query<br />

query


Users<br />

1<br />

Analysis tools writer SpreadSheet D<strong>at</strong>am<strong>in</strong>ig<br />

2<br />

M_d<strong>at</strong>abase<br />

MOLAP/ROLAP server<br />

3<br />

Metad<strong>at</strong>a<br />

D<strong>at</strong>a Warehouse<br />

D<strong>at</strong>a<br />

Mart<br />

Integr<strong>at</strong>or<br />

Wrapper Wrapper Wrapper<br />

4<br />

IMS<br />

RDMS<br />

Others


System Architecture<br />

Slice 1<br />

Appl. 1<br />

Java<br />

Appl. 2<br />

PB<br />

Appl. 3<br />

Delphi<br />

Consulte rules and<br />

libraries<br />

Slice N<br />

Oracle<br />

Sybase<br />

DB/2<br />

Appl. n<br />

MATLA<br />

N-2 slices<br />

Informix


Some Biological problems<br />

• Modific<strong>at</strong>ion of Sugar Cane, Eucalyptus<br />

and chickens<br />

• Cancer diagnostics<br />

• Drug performance aga<strong>in</strong>st HIV and Malaria<br />

• Understand<strong>in</strong>g of the cell division cycle<br />

• Reconstitution of nervous tissue<br />

• Design of drugs<br />

• Prediction of new virus


Formaliz<strong>in</strong>g Biological Problems


Σ<br />

Wh<strong>at</strong> genes regul<strong>at</strong>e the<br />

p<strong>at</strong>hway A->B->C->D ?<br />

Wet Lab<br />

Proteome<br />

Transcriptome<br />

Genome<br />

P<strong>at</strong>hways


Choice of adequ<strong>at</strong>e cluster<strong>in</strong>g<br />

technique


Cluster<strong>in</strong>g<br />

measurement<br />

Class 2<br />

[high variance]<br />

measurement<br />

Class 1<br />

[low variance]


Looser clusters due<br />

to large variance<br />

Cluster<strong>in</strong>g


Design classifier<br />

v−erb−b2 avian erythroblastic leukemia viral oncogene homolog 2<br />

2<br />

1.5<br />

1<br />

0.5<br />

0<br />

−0.5<br />

−1<br />

−1.5<br />

LINEAR CLASSIFIER (DISPERSED−GAUSSIAN) w/ σ = 0.600<br />

the tumor sample from P<strong>at</strong>ient 20<br />

7−th (0.416607)<br />

BRCA1<br />

BRCA2/sporadic<br />

1−th (0.389593)<br />

−3.5 −3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1<br />

phosphofructok<strong>in</strong>ase, pl<strong>at</strong>elet


Dimensionality Reduction<br />

x2<br />

Wh<strong>at</strong> is the m<strong>in</strong>imum number<br />

of genes th<strong>at</strong> is enough to<br />

dist<strong>in</strong>guish two Biological<br />

st<strong>at</strong>es?<br />

x1


Filter Design


Applic<strong>at</strong>ion


<strong>Model<strong>in</strong>g</strong> Dynamical Systems


Cell<br />

peptide<br />

other signals


peptide<br />

other signals<br />

mRNA<br />

Transcr.<br />

Transl<strong>at</strong>.<br />

Prote<strong>in</strong>s<br />

Pool<br />

Metabolic<br />

P<strong>at</strong>hways


peptide<br />

other signals<br />

GENES<br />

NETWORK<br />

mRNA<br />

Transcr.<br />

Transl<strong>at</strong>.<br />

Prote<strong>in</strong>s<br />

Pool<br />

Metabolic<br />

P<strong>at</strong>hways


x1fp<br />

.<br />

.<br />

.<br />

x1<br />

Division Steps<br />

z<br />

u1<br />

y1fp<br />

x1f<br />

y1f<br />

I<br />

u2<br />

u5<br />

p<br />

w1fp<br />

vfp<br />

w2fp<br />

w1f<br />

v w2f<br />

y2fp<br />

w1<br />

w2<br />

x3f<br />

x4f<br />

y1f<br />

y2f<br />

y2f<br />

y1<br />

y2<br />

z<br />

Forward Signal<br />

Feedback to p<br />

Feedback to previous layer<br />

.<br />

.<br />

.<br />

x6fp<br />

x6f<br />

x6


Knockout<br />

x1fp<br />

.<br />

.<br />

.<br />

x1<br />

Division Steps<br />

z<br />

u1<br />

y1fp<br />

x1f<br />

y1f<br />

I<br />

u2<br />

u5<br />

p<br />

w1fp<br />

vfp<br />

w2fp<br />

w1f<br />

v w2f<br />

y2fp<br />

w1<br />

w2<br />

x3f<br />

x4f<br />

y1f<br />

y2f<br />

y2f<br />

y1<br />

y2<br />

z<br />

Forward Signal<br />

Feedback to p<br />

Feedback to previous layer<br />

.<br />

.<br />

.<br />

x6fp<br />

x6f<br />

x6


System identific<strong>at</strong>ion


peptide<br />

Cell<br />

other signals<br />

GENES<br />

NETWORK<br />

mRNA<br />

Transcr.<br />

Transl<strong>at</strong>.<br />

Prote<strong>in</strong>s<br />

Pool<br />

Metabolic<br />

P<strong>at</strong>hways<br />

microarray<br />

proteome


F<strong>in</strong>d the architecture of a gene regul<strong>at</strong>ion network from<br />

microarray d<strong>at</strong>a.<br />

NMSE 2<br />

Predictor 2<br />

NMSE 1<br />

Target Gene Predictor 1<br />

NMSE 3<br />

Predictor 3


System dynamics simple


System identified


A family of <strong>M<strong>at</strong>hem<strong>at</strong>ical</strong><br />

problems


Design of classifier, filter or<br />

dynamical system<br />

Optimiz<strong>at</strong>ion<br />

Comb<strong>in</strong><strong>at</strong>ory<br />

or cont<strong>in</strong>uos<br />

St<strong>at</strong>istical<br />

Estim<strong>at</strong>ion<br />

Algebraic<br />

Represent<strong>at</strong>ion


The constra<strong>in</strong>ed estim<strong>at</strong>ion problem<br />

Error<br />

ψ opt<br />

Ψ<br />

Sample size


Other areas with similar<br />

problems<br />

• F<strong>in</strong>ances<br />

• Market<strong>in</strong>g<br />

• Digital TV<br />

• Petrol Industry<br />

• Neuro Sciences<br />

• ...

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!