IkagakukennkyuuStatGenet2011 - Statistical Genetics, Kyoto

Download Report

Transcript IkagakukennkyuuStatGenet2011 - Statistical Genetics, Kyoto

Statistical Genetics
統計遺伝学
2011/06/06
Ryo Yamada
Unit of Statistical Genetics
Center for Genomic Medicine
Graduate School of Medicine
Kyoto University
太陽
おけら
てのひら
みみにみみず
あめんぼ
私が両手をひろげても、
お空はちっとも飛べないが、
飛べる小鳥は私のやうに、
地面(じべた)を速くは走れない。
私がからだをゆすっても、
きれいな音は出ないけど、
あの鳴る鈴は私のやうに、
たくさんな唄は知らないよ。
みつばち
鈴と、小鳥と、それから私、
みんなちがって、みんないい。
とんぼ
かえる
瓜二つ
瓜の蔓に茄子はならぬ
鳶が鷹を生む
カエルの子はカエル
What to study?
How to study?
What do you want to know?
How do you want to know it?
Genetics
• Genotype
• Phenotype
Genetics
• Genotype
• Identity
• Phenotype
• Variation
How to grab “Genotype” and “Phenotype” with their “Identity” and “Variation”.
One way is to make a catalogue of facts among “them”.
The other way is to give a strategy to make the catalogue.
Genotype
Phenotype
Intermediate
phenotype
Terminal
phenotype
Graph (Theory)
グラフ(理論)
Pedigree 家系図
Pedigree
家系図
phylogenetic tree
系統樹
Phylogety, a tree
Distance 距離
Distance: More than one definition
距離にもいろいろな定義がある
Make graph from distance info
距離情報からグラフを作る
Distance is a way to quantitate relation.
Three items that make a triangle can be
connected as a star without changing their
pathway-distance.
Difference among graphs
グラフ間の違い
Same or different as a tree?
Difference among graphs
グラフ間の違い
Topology (Shape) and length of edges
位相(形)と辺の距離
Same or different as a tree?
Items can be expressed as a tree when
distance among them are given.
Different “distances” give different trees.
Graphs for data-analyses
グラフによるデータ解析
Heatmap
Data-mining approach also uses trees.
Tree needs “distance” with its definition.
Clustering methods also need definition to make tree structure..
Data give trees and change the order in items.
When the order in items are not changed and when relation
among the items are displayed, it is correlation matrix.
Original data
Relations among lines
Relations among columns
Relations among columns and lines
Pedigrees are NOT graphs
家系図はグラフでは ない
Trees in classical genetics:
Pedigrees
Relation between humans.
Relation between chromosomes.
A human relation can be
multiple chromosomal
relations.
Chromosomes have two
parental chromosomes.
But a base has only one
parental base.
Sexual and Asexual
Reproduction
Graph
有性生殖と無性生殖、とグラフ
Genotype
Phenotype
Intermediate
phenotype
Terminal
phenotype
Types of Data
データ タイプ
Categorical data and sets.
カテゴリ型データと集合
Ordered and Non-ordered.
High-dimensional data and graph
Genotype
Phenotype
Intermediate
phenotype
Terminal
phenotype
Networks
ネットワーク
DNA塩基配列
バリアント
DNA配列
エピゲノム修飾
?
次世代
eQTL
シークエンス
ネットワーク
?
?
?
(転写物・翻訳物)
?
GWAS
?
E1
D1aD1D1b
E2
E3
D2b
D2a D2 D2c
E4
D3
疾患に共通
する因子
E5
D4
D5
疾患とその
亜分類
Components of graphs
グラフの部品
Concepts of regulations/interactions
制御/相互関係の概念
Graphs are being used for biology
Networks are complex
ネットワークは複雑
We can not grab the graph as a whole at once
グラフ全体を一発で了解することは無理
Transition of Stata
状態推移
step-by-step
順番に
Markov-chain
マルコフ連鎖
Bayesian networks
ベイズネットワーク
Genetic Heterogeneity and Transition of Stata
遺伝的多様性と状態推移
Three components to
make genetic heterogeneity
遺伝的多様性を作る3要素
Mutations
Recombinations
Genetic drifts
変異
組換え
遺伝的浮動
Variants will drift out from the world.
疾患原因・薬剤応答性遺伝因子探索
GWAS
Complementary strand, for what?
Cross-overs and
recombinatios
What is the relation between
crossovers and identity of origin?
What is the distribution of segment-length between
crossovers?
Exponential...
Some chromosomes leave many
copies but others none.
Variants will drift out from the world.
Chromosomal relations in generations.
Population.
Chronological changes.
Spatial changes.
Time,Space
時間、空間
• Dimensions 次元
• Dimensions of data
Data analyses and space of data
Alleles and haplotypes and their relation.
RNA codon table
RNA codon table can be drawn as a tree.
Space may not be Euclidean
空間はユークリッド的でないかも
Closed space
閉じた空間
Populations in “Space” and “Time”.
Finite space vs. Infinite Space
Non-linear
非線形
Stable
Equilibrium
安定
定常
Models how to
handle space
and time.
• ? in Statistical genetics?
– Welcome
• Any questions on how to
handle bio-medical data ?
– Also welcome
Ryo Yamada, M.D., Ph.D.
Unit of Statisical Genetics
4F Kaibou-center building
phone: (+81)75-753-9470
fax: (+81)75-753-9284
[email protected]
統計遺伝学分野
山田 亮
医学部解剖センター棟4階
http://www.med.kyoto-u.ac.jp/E/grad_school/introduction/1525/