Transcript Data-intensive Computing: Case Study Area 1: Bioinformatics

Data-intensive Computing: Case
Study Area 1: Bioinformatics
B. Ramamurthy
4/11/2016
1
Human Genetics
•
•
•
•
•
•
Genomics
Human Genome project
Proteomics
Diseasome
Tree of life project
Phylogenetics
4/11/2016
2
Human cell
•
Base pair of DNA: CG, AT
– C – cytosine, G – guanine, A – adenine , T - thymine
•
•
•
•
•
•
•
•
Each human cell contains approximately 3 billion base pairs.
The DNA of a single cell contains so much information that if it were represented
in printed words, simply listing the first letter of each base would require over 1.5
million pages of text!
If laid end-to-end, the DNA strand measures about 2 – 3 meters.
DNA is a single large molecule at the nucleus of cell
It is coiled a double helix
Each strand of the DNA molecule is made of A, C, G and T: example:
AAAGTTCTTAATTA that will be matched on the other strand by the matching base:
TTTCAAGAATTAAT
These string of alphabets contain all the codes needed for the human functions
Ref text: Bioinformatics: Databases, tools and algorithms, by. O. Bosu and S.K.
Thukral
4/11/2016
3
More details
• Sequence of base pairs are grouped to make sense: genes
• When a gene inside needs to be activated, the DNA
molecule at the cell nucleus uncoils and unfurls to the right
extent to expose that gene
• From the exposed ends of the DNA a RNA is formed.
• mRNA or messenger RNA is formed that carries with it the
“print” of the open DNA section
• RNA and DNA differ in one respect: RNA does not contain T
or thymine but it has uracil (U). RNA is short-lived
• Once mRNA is formed open sections of the DNA close off.
4/11/2016
4
Protein formation
• mRNA travels to the cytoplasm where it meets the ribosome (rRNA)
• Ribosome reads the code in the mRNA (codon) and form the amino
acids.
• Twenty amino acids are prevalent in human cells. Ex: codon GCU
GCC GCA correspond to alanine
• In effect ribosome is a process control computer that takes in as
input codons and produces amino acids as output.
• Amino acids polymerize and form polypeptide chains called
proteins
• Proteins fold and form the basic structures such as skin and hair.
• Even though brain controls major human functions at the cell level
it is the DNA that has the command and control.
• DNA is fixed code for a given human. (WORM characteristics)
4/11/2016
5
Life’s processes
• DNA is “program” that controls functions,
operations and structure of a cell and in turn
that of our life processes.
• Life processes are in fact dependent of the
program in a DNA and the hundreds of
millions of ribosomes.
• Life in this context appears as an immense
distributed system.
4/11/2016
6
Bioinformatics
• Can we study, understand and analyze the complexity of the
immensely complex system? It structure and programs?
• University of Arizona’s tree of life project (ToL): http://tolweb.org
• Human Genome project (NIH and DOE): collecting approximately
30,000 genes in human DNA and determining the sequences three
billion bases that make up the human DNA.
• Out of the 30000 genes we do not know the functions of more than
50% of them.
• 99.9% of the nucleotide sequence is same for all of us
• 0.1% is attributed to individual differences such as race, color of
skin, disposition to diseases
• High throughput sequencing is generating ultra scale biological
data: how to analyze this data?
• That is a data-intensive problem.
4/11/2016
7
Existing solutions?
• Traditional databases: store, retrieve, analyze
and/or predict huge biological data
• Software tools for implementing algorithms,
and developing applications for in-silico
experiments
• Visualization tools, user interfaces, web
accessibility for search through data
• Machine learning and data mining
methodologies.
4/11/2016
8
Databases
•
•
•
•
•
•
•
•
•
•
Taxonomy DB
Genomics
Sequence db
Structure db
Proteomic database (PDB)
Micro-array db
Expression db
Enzyme db
Disease db
Molecular biology db
4/11/2016
9
Tools
• Data analysis tools
– MySQL
– Perl
• Prediction tools
– Clustering
• Modeling tools
– Surface prediction, predicting area of interest,
protein-protein interaction
• Alignment tools
4/11/2016
10
How can we help?
• How can we leverage our knowledge of large
scale data management to address
bioinformatics problems? DC methods.
• Large number of tools and data: how we
standardize the efforts so that they are
complementary or repetitive? Cloud
computing.
4/11/2016
11
Text Mining vs Genetic Sequence
Mining (Dot plot)
C
O
R
E
L
A
T
I
O
N
S
H
I
P
4/11/2016
R
R
E
L
A
T
I
O
N
S
A
C
T
C
T
A
G
G
A
G
T
C
G
A
T
A
A
T
T
C
G
A
T
C
12