Bioinformatics t9-t10-biocheminformatics v2014

FBW

9-12-2014

Wim Van Criekinge

Examen

<html>

<title>Examen Bioinformatica</title>

<center>

<head>

<script>

rnd.today=new Date();

rnd.seed=rnd.today.getTime();

function rnd() {

rnd.seed = (rnd.seed*9301+49297) % 233280;

return rnd.seed/(233280.0);

};

function rand(number) {

return Math.ceil(rnd()*number);

};

</SCRIPT>

</head>

<body bgcolor="#FFFFFF" text="#00FF00" link="#00FF00">

<script language="JavaScript">

document.write('<table>');

document.write('<tr>');

document.write('<td><a href="index.html" ><img border=0 src="' + rand(713) + '.jpg" width="520" height="360"></a></td>');

rand(98);


rand(98);


rand(98);


rand(98);

document.write('</tr>');

• The keywords can be

– genome structure

– gene-organisation

– known promoter regions

– known critical amino acid residues.

• Combination of functional

modelorganism knowledge

• Structure-function

• Identify similar areas of biology

• Identify orthologous pathways (might

have different endpoints)

Comparative Genomics: The biological Rosetta

Example: Agro

Known “lethal” genes

from worm, drosphila

Sequence Genome

Filter for drugability”,

tractibility & novelty

Example: Extremophiles

Known lipases

Filter for

“workable”lipases

at 90º C

Look for species

with interesting

phenotypes

Clone and produce in large quantities

Washing Powder additives

Sequence Genome

Functional Foods

Convert Highly Energetic Monosaccharides to Dextrane

Drug Discovery: Design new drugs by computer ?

Problem: pipeline cost rise linear, NCE steady

Money: bypassing difficult, work on attrition

Every step requires specific computational tools

• Drugs are generally defined as molecules which affect biological processes.

• In order to be effective, the molecule must be present in the body at an adequate concentration for it to act at the specific site in the body where it can exert its effect.

• Additionally, the molecule must be safe -- that is, metabolized and eliminated from the body without causing injury.

• Assumption: next 50 years still a big market in small chemical entities which can be administered orally in form of a pill (in contrast to antibodies) or gene therapy …

Drug Discovery: What is a drug ?

• Taxol a drug which is an unmodified natural compound, is the exception

• Most drugs require “work” -> need for target driven pipeline

• Humane genome is available so all target are identified

• How to validate (within a given disease area) ?

http://www.bris.ac.uk/Depts/Chemistry/MOTM/taxol/taxol.pdb

http://www.bris.ac.uk/Depts/Chemistry/MOTM/taxol/taxol.pdb

• target - a molecule (often a protein) that is instrumental

to a disease process (though not necessarily directly

involved), which may be targeted with a potential

therapeutic.

• target identification - identifying a molecule (often a

protein) that is instrumental to a disease process (though

not necessarily directly involved), with the intention of

finding a way to regulate that molecule's activity for

therapeutic purposes.

• target validation - a crucial step in the drug

development process. Following the identification of a

potential disease target, target validation verifies that a

drug that specifically acts on the target can have a

significant therapeutic benefit in the treatment of a given

disease.

Drug Discovery: What is a target ?

Phenotypic Gap

# genes with

known function

Total # genes

Number of genes

1980 1990 2000 2010

Functional Genomics ?

More than running chip experiments !

Proposal to prioritize

hypothetical protein

without annotation, nice

for bioinformatics and

biologist

“Optimal” drug target

Predict side effect

Where is optimal drug target ?

How to correct disease state

Side effects ?

Genome-wide RNAi

RNAI vector

bacteria producing ds RNA for

each of the 20.000 genes

proprietary nematode

responding to RNAi

20.000 responses

20.000 genes insert

library

Normal insulin signaling

Reduced insulin signaling

fat storage LOW

fat storage HIGH

Type-II Diabetes

20,000 bacteria

each containing

selected

C. elegans gene

select genes with desired phenotypes

proprietary C.elegans strains

• sensitized to silencing

• sensitized to relevant pathway

Industrialized knock-downs

Pharma is conservative

Molecular functions of 26 383 human genes

Structural Genomics

Lipinsky for the target ?

Database of all “drugable” human genes

Drug Discovery: Design new drugs by computer ?

screening - the automated examination and

testing of libraries of synthetic and/or organic

compounds and extracts to identify potential drug

leads, based on the compound's binding affinity

for a target molecule.

screening library - a large collection of

compounds with different chemical properties or

shapes, generated either by combinatorial

chemistry or some other process or by collecting

samples with interesting biological properties.

High Throughput Screening: Quick and Dirty…

from 5000 compounds per day

Drug Discovery: Screening definitions

• At the beginning of the 1990s, when the

term "high-throughput screening" was

coined, a department of 20 would

typically be able to screen around 1.5

million samples in a year, each

researcher handling around 75,000

samples. Today, four researchers using

fully automated robotic technology can

screen 50,000 samples a day, or around

2.5 million samples each year.

Drug Discovery: Screening Throughput

Robotic arm

Read-out

Fluorescence /

luminescence

Distribution

96 / 384 wells

Optical Bank

for stability

Drug Discovery: HTS – The Wet Lab

• Available molecules collections from pharma,

chemical and agro industry, also from

academics (Eastern Europe)

• Natural products from fungi, algae, exotic

plants, Chinese and ethnobotanic medicines

• Combinatorial chemistry: it is the generation

of large numbers of diverse chemical

compounds (a library) for use in screening

assays against disease target molecules.

• Computer drug design (from model

substrates or X-ray structure)

Drug Discovery: Chemistry Sources

Drug Discovery

HIT LEAD

• initial screen established

• Compounds screened

• IC50s established

• Structures verified

• Minimum of three independent

chemical series to evaluate

• Positive in silico PK data

Drug Discovery: HIT

• When the structure of the target is unknown,

the activity data can be used to construct a

pharmacophore model for the positioning of

key features like hydrogen-bonding and

hydrophobic groups.

• Such a model can be used as a template to

select the most promising candidates from the

library.

Drug Discovery: Hit/lead computational approaches

• lead compound - a potential drug candidate emerging from a screening process of a large library of compounds.

• It basically affects specifically a biological process. Mechanism of activity (reversible/ irreversible, kinetics) established

• Its is effective at a low concentration: usually nanomolar activity

• It is not toxic to live cells

• It has been shown to have some in vivo activity

• It is chemically feasible. Specificity of key compound(s) fromeach lead series against selected number of receptors/enzymes

• Preliminary PK in vivo (rodent) to establish benchmark for in vitro SAR

• In vitro PK data good predictor for in vivo activity

• Its is of course New and Original.

Drug Discovery: Lead ?

Christopher A. Lipinski, Franco Lombardo, Beryl W. Dominy, Paul J. Feeney

"Experimental and computational approaches to estimate solubility and

permeability in drug discovery and development settings":

"In the USAN set we found that the sum of Ns and Os in the molecular formula was

greater than 10 in 12% of the compounds. Eleven percent of compounds had

a MWT of over 500. Ten percent of compounds had a CLogP larger than 5 (or

an MLogP larger than 4.15) and in 8% of compounds the sum of OHs and NHs

in the chemical structure was larger than 5. The "rule of 5" states that: poor

absorption or permeation is more likely when:

A. There are less than 5 H-bond donors (expressed as the sum of OHs and

NHs);

B. The MWT is less than 500;

C. The LogP is less than 5 (or MLogP is < 4.15);

D. There are less than 10 H-bond acceptors (expressed as the sum of Ns and

Os).

Compound classes that are substrates for biological transporters are exceptions to

the rule."

Lipinski: « rule of 5 »

• A quick sketch with ChemDraw, conversion to a

3D structure with Chem3D, and processing by

QuikProp, reveals that the problem appears to be

poor cell permeability for this relatively polar

molecule, with predicted PCaco and PMDCK

values near 10 nm/s.

• Free alternative (Chemsketch / PreADME)

(Celebrex)

Methyl in this position makes it a weaker cox-2 inhibitor,

but site of metabolic oxidation and ensures an acceptable clearance

Drug-like-ness

To assist combinatorial chemistry, buy specific compunds

Structural Descriptors: (15 descriptors)

Molecular Formula, Molecular Weight, Formal Charge, The Number of Rotatable Bonds, The Number of Rigid Bonds, The Number of Rings, The Number of Aromatic Rings, The Number of H Bond Acceptors, The Number of H Bond Donors, The Number of (+) Charged Groups, The Number of (-) Charged Groups, No. single, double, triple, aromatic bonds

Topological Descriptors:(350 descriptors)

• Topological descriptors on the adjustancy and distance matrix

• Count descriptors

• Kier & Hall molecular connectivity Indices

• Kier Shape Indices

• Galvez topological charge Indices

• Narumi topological index

• Autocorrelation descriptor of atomic masses, atomic polarizability, Pauling electronegativity and van der Waals radius

• Information content descriptors

• Electrotopological state index (E-state)

• Atomic-Level-Based AI topological descriptors

Physicochemical Descriptor:(10 descriptors)

AlogP98 (calculated logP), SKlogP (calculated logP), SKlogS in pure water (calculated water solubility), SKlogS in buffer system (calculated water solubility),SK vap (calculated vapor pressure), SK bp (calculated boiling point), SK mp (calculated meling point), AMR (calculated molecular refractivity), APOL(calculated polarizability), Water Solvation Free Energy

Geometrical Descriptor:(9 descriptors)

Topological Polar Surface Area, 2D van der Waals Volume, 2D van der Waals Surface Area, 2D van der Waals Hydrophobic Surface Area, 2D van der Waals Polar Surface Area, 2D van der Waals H-bond Acceptor Surface Area, 2D van der Waals H-bond Donor Surface Area, 2D van der Waals (+) Charged Groups Surface Area, 2D van der Waals (-) Charged Groups Surface Area

• What can you do with these descriptors ?

• Cluster entire chemical library

– Diversity set

– Focused set

Drug Discovery: Hit/lead computational approaches

• Structure is known, virtual screening -> docking

• Many different approaches– DOCK

– FlexX

– Glide

– GOLD

• Including conformational sampling of the ligand

• Problem: – host flexibility

– solvatation

• Example: Bissantz et al.– Hit rate of 10% for single scoring function

– Up to 70% with triple scoring (bagging)

Drug Discovery: Docking

• Given the target site:

• Docking + structure generator

• Specialized approach: growing substituent on a core– LUDI

– SPROUT

– BOMB (biochemical and organic model builder)

– SYNOPSIS

• Problem is the scoring function which is different for every protein class

Drug Discovery: De novo design / rational drug design

Drug Discovery: Novel strategies using bio/cheminformatics

- HTS ? Chemical space is big (1041)

- Biased sets/focussed libraries -> bioinformatics !!!

- How ? Use phylogenetics and known structures to define

accesible (conserved) functional implicated residues to

define small molecule pharmacophores (minimal

requirements)

- Desciptor search (cheminformatics) to construct/select

biased compound set

- ensure serendipity by iterative screening of these

predesigned sets

Drug Discovery

Toxigenomics

Metabogenomics

• Preclinical - An early phase of development

including initial safety assessment

Phase I - Evaluation of clinical pharmacology,

usually conducted in volunteers

Phase II - Determination of dose and initial

evaluation of efficacy, conducted in a small

number of patients

Phase III - Large comparative study

(compound versus placebo and/or established

treatment) in patients to establish clinical

benefit and safety

Phase IV - Post marketing study

Drug Discovery: Clinical studies

Drug Discovery & Development: IND filing

Hapmap

Pharmacogenomics

Predictive/preventive – systems biology

Sneak preview

Bioinformatics (re)loaded

Sneak preview

Bioinformatics (re)loaded

• Relational datamodels

– BioSQL (MySQL)

• Data Visualisation

– Interface

• Apache

• PHP

• Large Scale Statistics

– Using R

Bioinformatics t9-t10-biocheminformatics v2014

Education