slides from M. Pellegrini
Types of Studies
synthetic lethals: pairs of gene knockouts
drug screen: gene knockouts vs. chemicals
RNAi: easy way to knock down target
genes in human cell lines etc.
First vs Second
Today: Our Last Class
Practice Final and solutions posted
Summary: a recipe for genomics
course evaluations & questionaire.
From Genomics to
Includes slides from M. Pellegrini
HW6 due today.
assembly algorithm reading (PS3) will be
handed out on Wednesday.
next Monday is a holiday!
HW7 will be posted Wednesday, not due till 2
weeks from today (June. 3).
RNA-seq and Alternative
example of what we can get from
transcriptional analysis: yeast metabolic
the new frontier: ENCODE analysis of the
Yeast Metabolic Cycle
Slides from M. Pellegrini
HW5 due today
HW6 posted today: gene expression
analysis using R statistics package
midterm key posted today
midterms will be returned in discussion
using DNA Microarrays
Includes slides from Matteo Pelle
Genome-wide Analysis of
and DNA Methylation
Includes slides from M.
DNA within nuclei is
compacted many-fold so that
several meters of DNA can t
into a one micron nucleus
histones are the core proteins
Say you have generated a very large dataset of short read sequences
from a soil sample, and assembled these reads using a process similar
to the Sargasso Sea shotgun sequence assembly described in lecture.
The resulting contigs will h
Phenotype criteria for gene interaction studies
Say a colleague proposes to do a genetic interaction study based on a
specic phenotype. Briey explain what criteria the phenotype must
meet in order for this to be feasible, if any (can this be done for just
Graduate Term Paper
option: choose an existing genomics study
(paper) and critique its methods / results
using principles from this class.
option: propose a new kind of genomics
study by applying these principles.
option: perform a brief analysis of an
Proteomics Pipeline Steps
Outline the basic experimental steps required for identifying a protein
from a sample using Mass Spec (MS). Can a single MS spectrum
identify all or a large proportion of the proteins in a tissue?
Proteomics Pipeline Steps Answer
Yeast Metabolic Cycle Pathway Identication
The yeast metabolic cycle study identied a large cluster of
genes up-regulated during Ox phase.
Assume we have pathway annotations for all yeast genes that
assign them to specic biological pathways.
How would you
slides from Matteo Pellegrini, MCDB
Basic Local Alignment Sequence Tool
Uses shortcut to compute alignments of a
sequence against a database very quickly
Typically takes about a minute to align a
sequence against a database of one million
Midterm, HW etc.
Midterms not yet graded; will be handed back
HW5 posted yesterday, due next week.
Starting today we focus on different genomics
application areas, to understand the biological
questions they answer, and to see how basic
Expression Analysis by RNAseq?
Traditionally, gene expression studies used microarrays to measure
the transcript levels of all genes in a sample simultaneously. Can
RNAseq data yield this same measurement of all transcript levels? If
so, briey explain how
HW1 due here or by email
HW2 will be on the class website.
Today we discuss Sequence variation,
Wednesday well go over basic statistics for
Two short chapters for next lecture.
More concept tests after todays lecture.
Testing Predicted Regulatory Sites
Say you have analyzed ChIP-chip experiments for a set of
transcription factors, using a range of conditions where those
transcription factors should be active. For each transcription factor,
you identied a sequence motif
Tumor Type Signatures
Within a general category of cancer (e.g. lung cancer) there are tumor
subtypes that differ greatly in their treatment and prognosis.
Pathologists specialize in identifying the tumor type from a patient
biopsy (classically, by examin
Say you are given pairwise co-function p-values for 4000 bacterial
genes. State how you would lter the set of predictions to predict
genes that function together while keeping the error rate to fewer than
0.05 false positives (on av
Midterm Next Monday
covers up to last week and HW4
mix of concept questions and problems to
solve (e.g. calculate a probability)
no questions on Python or procedural
details from the homework.
homework keys posted
practice questions will be posted tod
Statistical Inference & Hypothesis Testing
Statistical Inference & Hypothesis Testing
Probability and statistics are essential for answering genomics
questions, because of the huge size of genomics datasets.
Even a small rate of error ca
Scoring and Dynamic Programming
A structural biologist proposes two alternative scoring systems for
aligning a protein sequence to any known protein structure.
environment classes: a list of environment classes is dened
(e.g. fully solvent-exposed; hydrop
Transcription Factor Binding and Nucleosomes
Transcription factor binding sites are characterized as sequence motifs
(typically displayed as a logo diagram). Often these are low in
specicity, e.g. with only a few absolutely conserved nucleotides, there
Measuring Random vs.
random errors can be broadly categorized
as problems of insufcient sampling.
Need to test whether our tree is likely to be
quite different if we had a complete sample.
We also must consider the possibility of
Chemistry 100 / 200
Goal: infer past history that produced a set of
modern characters (sequences, typically).
Characters: e.g. sequence differences
UPGMA is a general-purpose clustering algorithm that can be applied
to inferring the family tree structure of a set of related sequences. How
general is this algorithm? I.e. do you expect it to work on all kinds of
Selection Pressure Metrics:
Downloading Evolutions Experiments
Christopher Lee, UCLA Dept. of Chemistry & Biochemistry
Define a simple measure of selection
Example: HIV drug resistance evolution
Extending selection metrics to the big
The Monty Hall Problem
Imagine youre a contestant on the Monty Hall Show. There are
three doors, one concealing a valuable prize, and if you pick the
right door, you win. But theres a twist: after you say which door
youve picked, Monty always opens anothe