Presentation is loading. Please wait.

Presentation is loading. Please wait.

Classification of tissues and samples 指導老師:藍清隆 演講者:張許恩、王人禾.

Similar presentations


Presentation on theme: "Classification of tissues and samples 指導老師:藍清隆 演講者:張許恩、王人禾."— Presentation transcript:

1 Classification of tissues and samples 指導老師:藍清隆 演講者:張許恩、王人禾

2 introduction The use of microarrays to find group of genes that can be used diagnostically to determine the disease that an individual is suffering from, or prognostically to predict the success of a course of therapy or results of an experiment. Purpose: find a small number of genes that can predict to which group each individual belongs

3 Data set 9A Bone marrow samples are taken from 27 patients suffering from acute lymphoblastic leukemia(ALL) and 11 patients suffering from acute myeloid leukemia(AML)

4 Method of classification Describe methods that allow you to predict the class to which an individual belongs, based on gene expression measurements Assume that we have already selected a small number of genes whose expression measurements we use, and are not using all the genes on the microarray

5 Two concept central to classification Separability and linearity

6 Separability Separable: the different groups to which the samples belong occupy different regions of the gene expression space Non-Separable: the different groups to which the samples belong are mixed together in the same region of gene expression space

7 linearity Linearly separable: possible to partition the space between the two (or more) group using straight lines Non- Linearly separable: separable, but are not possible to partition the groups using straight lines

8

9

10

11 Method of classification 1.K-nearest neighbours 2.Nearest centroid 3.Linear discriminant analysis 4.Neural networks 5.Support vector machines

12 K-nearest neighbours Steps: 1.We look at the gene expression measurements for the sample we are trying to classify 2.Find the nearest of the known samples as measured by an appropriate distance measure 3.The class of the sample is the class of the nearest sample

13

14 K-nearest neighbours 2 parameter: k & l

15 Centroid classification Steps: 1.For each class, calculate the center of mass of the points of the representative samples 2.Calculate the distance between the position of the sample to be classified and each of the centers of mass of the classes using an appropriate distance measure 3.Assign the sample to the class whose center of mass is nearest to it

16

17 Centroid classification

18 Linear discriminant analysis Steps: 1.Calculate a straight line (in two dimensions) or hyperplane (in more than two dimensions) that separates two known classes so as to minimise the within class variance on either side of the line and maximise the between class variance(figure 9.4) 2.The class of the unknown sample is determined by the side of the hyperplane on which the sample lies.

19

20 Linear discriminant analysis

21 Neural networks Steps: 1.Train the neural network using the samples with known classes 2.Apply the neural network to the new individual to determine it’s class

22

23 Neural networks

24 Support vector machines Steps: 1.Project the data from the known classes into a suitable high-dimensional space 2.Identify a hyperplane that separates two classes 3.The class of the new individual is determined by the side of hyperplane on which the sample lines

25 Support vector machines

26 謝謝大家 的支持!

27


Download ppt "Classification of tissues and samples 指導老師:藍清隆 演講者:張許恩、王人禾."

Similar presentations


Ads by Google