Bioinformatics to predict bacterial phenotypes - NMPDR

An Introduction to
“Bioinformatics to Predict
Bacterial Phenotypes”

Jerry H. Kavouras, Ph.D.

Lewis University

Romeoville, IL

Why study bioinformatics?

It’s a rapidly growing field in biology

and computer science that analyzes the

genetic information stored in DNA.


What are phenotypes?

Phenotypes are the observed manifestations

of genes, which are linear sequences of

nucleotides on chromosomes and plasmids.



Examples:



Cell shapes (bacillus, coccus)



Metabolic capacity (glucokinase)



Antibiotic resistance (
b
-
lactamase)

How are phenotypes determined?

Traditionally, microbiologists determine

phenotypes by culture methods and

microscopic techniques.


Examples:


Test



Phenotype


Gram stain


Cell wall


Catalase test


Degrades H
2
O
2



What phenotypes do you observe
in this Gram stain preparation?



Phenotypic observations

Bacillus

Length is 3
m
m

Gram negative

1
m
m

Gram positive

Coccus

Homologous genes

Homologous genes share similar sequences



--

Related by evolution



--

Good chance have similar function


Similar sequences are likely to be versions of

the same gene, or at least similar genes,

meaning there is a very good chance the

PHENOTYPE is present.

Use bioinformatics to predict
cellular phenotypes!

Possible to predict cellular phenotypes

from a linear sequence of chromosomal or

plasmid DNA if
the gene is already known.


Why is this important?

This means microbiologists can determine

the phenotypes for a microbe without

culture methods and microscopy.

Why use bioinformatics instead of
standard methods?

Microbiologists estimate the overwhelming

majority of bacteria in nature have not been

cultured (based on analyses of 16S rDNA

sequences)


Microbiologists can predict the phenotypes
of
UNCULTURED

bacteria!



--

Very powerful tool

Genomes & Bioinformatics

The number of complete genome sequences for

organisms is increasing steadily.


Someday microbiologists may rely on DNA

sequences for determining cell phenotypes.


This does not mean culture methods and

microscopy will not be important!

Real world applications

How can the same species of bacteria be

pathogenic and nonpathogenic?



Answer: Different strains!


Example:



Escherichia coli

K
-
12 (nonpathogenic)



Escherichia coli

O157:H7 (pathogen)

Real world applications (cont’d)

Because the chromosomal and plasmid

sequences are known for these strains,

bioinformatics can determine differences in

the set of genes (genome) for each organism.


Now use the Signature Genes tool at NMPDR to

actually answer this question.

Compare whole genomes

•
Go to
www.nmpdr.org

•
Select the Signature
Genes tool from the
left navigation bar

•
Choose
E. coli

O157:H7
as the reference

•
Choose another strain
of the O157:H7
serotype, strain
EDL933, in set 1

•
Choose two strains of
avirulent
E. coli

K12 in
the exclusion set 2

•
Click

Go

Find genes responsible for
virulence phenotype

•
847 genes common to two strains of
O157:H7 and absent from two lab strains
of
E. coli

•
Most of these are phage
-
associated

•
Included are 6 toxin genes

•
What might the remaining 841 genes be
doing?

How do you explore physical &
biological context of virulence
genes in
E. coli
?

1.
Study phylogenetic profiles



--

evolutionary relationships


among genes

2.
Study protein families



--

functional (analogous)


relationships among genes


Acknowledgements

Thanks to Leslie McNeil for her additions,

assistance, and comments in preparing this

exercise introduction.