Applications of Data Mining and Machine Learning in Bioinformatics

1
Applications of Data Mining and Machine Learning
in Bioinformatics

• Yen-Jen Oyang
• Dept. of Computer Science and Information
  Engineering

2
Basics of Protein Structures

• A typical protein consists of hundreds to
  thousands of amino acids.
• There are 20 basic amino acids, each of which is
  denoted by one English character.

3
20 Amino Acid - 1
Name Symbol Mass (-H2O) Side Chain Occurence ()
Alanine A, Ala 71.079 CH3- 7.49
Arginine R, Arg 156.188 HNC(NH2)-NH-(CH2)3- 5.22
Asparagine N, Asn 114.104 H2N-CO-CH2- 4.53
Aspartic acid D, Asp 115.089 HOOC-CH2- 5.22
Cysteine C, Cys 103.145 HS-CH2- 1.82
Glutamine Q, Gln 128.131 H2N-CO-(CH2)2- 4.11
Glutamic acid E, Glu 129.116 HOOC-(CH2)2- 6.26
Source http//prowl.rockefeller.edu/aainfo/struct
.htm

4
20 Amino Acid - 2
Name Symbol Mass (-H2O) Side Chain Occurence ()
Glycine G, Gly 57.052 H- 7.10
Histidine H, His 137.141 NCH-NH-CHC-CH2- __________ 2.23
Isoleucine I, Ile 113.160 CH3-CH2-CH(CH3)- 5.45
Leucine L, Leu 113.160 (CH3)2-CH-CH2- 9.06
Lysine K, Lys 128.17 H2N-(CH2)4- 5.82
Methionine M, Met 131.199 CH3-S-(CH2)2- 2.27
Phenylalanine F, Phe 147.177 Phenyl-CH2- 3.91
Proline P, Pro 97.117 -N-(CH2)3-CH- _________ 5.12

Source http//prowl.rockefeller.edu/aainfo/struct
.htm
5
20 Amino Acid - 3
Name Symbol Mass (-H2O) Side Chain Occurence ()
Serine S, Ser 87.078 HO-CH2- 7.34
Threonine T, Thr 101.105 CH3-CH(OH)- 5.96
Tryptophan W, Trp 186.213 Phenyl-NH-CHC-CH2- ___________ 1.32
Tyrosine Y, Tyr 163.176 4-OH-Phenyl-CH2- 3.25
Valine V, Val 99.133 CH3-CH(CH2)- 6.48

Source http//prowl.rockefeller.edu/aainfo/struct
.htm
6
Three-dimensional Structure of Myoglobin
Source Lectures of BioInfo by yukijuan
7
Prediction of Protein Functions

• Given a protein sequence, biochemists are
  interested in its functions and its tertiary
  structure.

8
Protein Classification Based on the Homology Model

• The sizes of modern protein databases are growing
  at fast rates.
• In order to expedite the process to identify
  protein functions, it is desirable to classify
  the concerned protein, before biochemistry
  experiments are conducted.

9

• One widely used approach to classify proteins is
  based on the homology model, i.e. classify
  proteins based on the similarities of amino acid
  sequences.
• BLAST and FASTA are two most widely used software
  utilities for computing the similarity between
  two sequences.
• We can cluster the proteins in an existing
  protein database in advance as the next slide
  exemplifies.

10
An Example of Similar Protein Sequences

• 3BP2_HUMAN MAAEEMHWPVPMKAIGAQNLLTMPGGVAKAGYLHKKGGT
  QLQLLKWPLRFVIIHKRCVYYFKSSTSASPQGAFSLSGYNRVMRAAEETT
  SNNVFPFKIIHISKKHRTWFFSASSEEERKSWMALLRREIGHFHEKKDLP
  LDTSDSSSDTDSFYGAVERPVDISLSPYPTDNEDYEHDDEDDSYLEPDSP
  EPGRLEDALMHPPAYPPPPVPTPRKPAFSDMPRAHSFTSKGPGPLLPPPP
  PKHGLPDVGLAAEDSKRDPLCPRRAEPCPRVPATPRRMSDPPLSTMPTAP
  GLRKPPCFRESASPSPEPWTPGHGACSTSSAAIMATATSRNCDKLKSFHL
  SPRGPPTSEPPPVPANKPKFLKIAEEDPPREAAMPGLFVPPVAPRPPALK
  LPVPEAMARPAVLPRPEKPQLPHLQRSPPDGQSFRSFSFEKPRQPSQADT
  GGDDSDEDYEKVPLPNSVFVNTTESCEVERLFKATSPRGEPQDGLYCIRN
  SSTKSGKVLVVWDETSNKVRNYRIFEKDSKFYLEGEVLFVSVGSMVEHYH
  THVLPSHQSLLLRHPYGYTGPR
• 3BP2_MOUSE MAAEEMQWPVPMKAIGAQNLLTMPGGVAKAGYLHKKGGT
  QLQLLKWPLRFVIIHKRCIYYFKSSTSASPQGAFSLSGYNRVMRAAEETT
  SNNVFPFKIIHISKKHRTWFFSASSEDERKSWMAFVRREIGHFHEKKELP
  LDTSDSSSDTDSFYGAVERPIDISLSSYPMDNEDYEHEDEDDSYLEPDSP
  GPMKLEDALTYPPAYPPPPVPVPRKPAFSDLPRAHSFTSKSPSPLLPPPP
  PKRGLPDTGSAPEDAKDALGLRRVEPGLRVPATPRRMSDPPMSNVPTVPN
  LRKHPCFRDSVNPGLEPWTPGHGTSSVSSSTTMAVATSRNCDKLKSFHLS
  SRGPPTSEPPPVPANKPKFLKIAEEPSPREAAKFAPVPPVAPRPPVQKMP
  MPEATVRPAVLPRPENTPLPHLQRSPPDGQSFRGFSFEKARQPSQADTGE
  EDSDEDYEKVPLPNSVFVNTTESCEVERLFKATDPRGEPQDGLYCIRNSS
  TKSGKVLVVWDESSNKVRNYRIFEKDSKFYLEGEVLFASVGSMVEHYHTH
  VLPSHQSLLLRHPYGYAGPR

11

• When a protein with unknown functions is
  inputted, the classification software identifies
  the protein clusters that contain most similar
  proteins.
• The biochemists then can predict the functions of
  the protein based on the output of the
  classification software.
• The protein clustering conducted in advance
  expedites the search process.

12
Applications of Data Classification in Microarray
Data Analysis

• In microarray data analysis, data classification
  is employed to predict the class of a new sample
  based on the existing samples with known class.

13

• For example, in the Leukemia data set, there are
  72 samples and 7129 genes.
• 25 Acute Myeloid Leukemia(AML) samples.
• 38 B-cell Acute Lymphoblastic Leukemia samples.
• 9 T-cell Acute Lymphoblastic Leukemia samples.

14
Model of Microarray Data Sets
15
Applications of Data Clustering in Microarray
Data Analysis

• Data clustering has been employed in microarray
  data analysis for
• identifying the genes with similar expressions
• identifying the subtypes of samples.