QBIO2001 Flashcards
Small scale data
1
Q
What is synthetic biology?
A
Synthetic biology- the use of molecular biology tools and techniques to forward engineer cellular behavior
2
Q
What is the design of the synthetic biology process?
A
- Design objectives and specifications:
a. Inputs and outputs
b. System performance - Design according to spec:
a. Conceptual design
b. Detailed design - System models composed from parts
a. Data may come from standardized database of biological parts - In silico verification
a. Analyse models
b. Simulate/predict behavior - Implementation
a. DNA assembly
b. DNA synthesis - Evolution
- Testing and characterization of the system
3
Q
What are 2 different types of design?
A
- Parts design
* Model based design
4
Q
What can synthetic biology be used for?
A
- Autoregulatory circuits
- Toggle switch
- Edge detection circuits
- Recombinase-based logic
5
Q
What is liquid chromotography and mass spectrometry used for?
A
Used for protein and metabolite analysis
6
Q
Describe the scientific method?
A
- Initial observation
a. Basic observations of the dataset
b. Observation allows people to generate a theory or consult a theory - Consult/Generate theory
- Generate hypothesis
a. Needs to be a testable statement by experiment
b. Needs to be a falsifiable by experiment
c. Needs to be very clear
d. This is the stage where variables are identified
i. Includes outcome variable
ii. Includes independent variable - Collect Data to test hypothesis
a. Measure the variables
b. From t test, can so whether hypothesis is right or not
c. Tests normally examine the null hypothesis
d. Low score on t test the alternative hypothesis is right
e. Have to watch out for confounding (potentially unexpected) variables - Analyse data
a. Graph data
b. Fit a model
c. If failed, another hypothesis can be generated
7
Q
What is an experimental unit?
A
Object of replication that can be assigned to a treatment
8
Q
What are examples of experimental unit?
A
- Could be an individual human or hundreds of mice could be 1 experimental unit
- For animal experiments, cage can be experimental unit (mice in the cage are put in the same conditions)
- Experimental unit could be regions of skin on one animal
- Individual cell in dish could be an experimental unit, as each cell has its own variability
9
Q
What is the between-subjects treatment group method?
A
- An experimental unit is chosen
- This experimental unit is put in one of three groups:
a. Experimental (treatment) group
i. Unknown change in dependent (outcome) variable
b. Negative control (untreated) group
i. No change in dependent variable expected
c. Positive control group
i. Determines test validity
ii. A known change in dependent (outcome) variable expected
iii. Exposed to a treatment that we know affects the dependent variables
10
Q
What are flaws in the between subjects-treatment groups method?
A
o Two major sources of variance in between subject design:
Variability of subjects/ experimental units (unsystematic variance)
Systematic variance -> treatment
o Need thousands of subjects to balance out variability
11
Q
What is the within subjects (paired/dependent) treatment groups method?
A
- Experimental unit is chosen
- That experimental unit goes through an initial test to determine baseline
- The same individual goes through the experimental treatment, and then a test
- The same individual also goes through the negative control (no treatment) after a period of time (to ensure the experimental treatment has worn off), and then a test
12
Q
What are flaws of the within subjects (paired/dependent) treatment groups method?
A
o Patient is not exactly the same as they were in initial test, as going through an experimental or negative control test may change their attitude towards the testing experience as a whole and, if the testing relies heavily on patient’s mindset, this might confound the results
Hence, the crossover design is used, where some experimental units go through experimental treatment first, then negative and vice-versa
o Time brings change, and since these experiments are done over time, all sorts of things can change over time
o People can drop out of the experiment, and then you only have half your data.
13
Q
What are advantages of the within subjects (paired/dependent) treatment groups method?
A
Number of experimental units is much less because variance is less
14
Q
What are two sources of variance?
A
- Unsystematic
- Systematic
15
Q
What is unsystematic variance?
A
Due to differences between experimental conditions (e.g. time of day, temperature, etc…) OR experimental units (e.g. genetics, sickness, etc…)
16
Q
What is systematic variance?
A
Due to the experimenter performing a treatment on all experimental units in one group but not those of another group (untreated negative control)
17
Q
What are experimental units representative of?
A
small sample (representative sample) of entire population
18
Q
What does randomization do and how can this be done?
A
• Randomisation -
o Experimental units assigned randomly to treatment groups to minimize unsystematic variation
Variance is equally/randomly distributed so it has minute influence on results
This can be done by computer algorithms
19
Q
What does blinding do?
A
o Eliminates bias that may increase variation
Eliminates psychological influences on outcome variable
20
Q
What is blinding?
A
when the patient doesn’t know which group they’ve been allocated to
21
Q
What is double blinding?
A
When both the patient and the researcher doesn’t know which group they’ve been allocated to: makes sure the researcher doesn’t give anything away/ doesn’t influence data collection
22
Q
What is blocking and what does it do?
A
o Similar number of experimental units assigned to each treatment group in a block to minimize effects of unavoidable variance
o Eliminates effect of known variance
o Make sure there are equal proportions of treatment/untreated groups across confounding variables
o Blocking used to spread the outcome variable over time
23
Q
What are power calculations?
A
Estimate the sample size required to detect an effect of a given size with a given degree of confidence
24
Q
What is the preliminary data needed for power calculations?
A
o Effect size- change in outcome variable you want to see
o Standard deviation of the outcome variable
o Significance level required (p<0.05)
o Type of statistical test
o Desired power (probability of detecting true effect)
25
Why are sample sizes constrained?
• Sample size constraints:
o Need to put request through ethics communities
o Funding
26
Why is graphing data useful?
• Graphing data is extremely useful in informing the researcher of the shape and distribution of their data
27
What is a big assumption in most tests?
• A big assumptions in most tests is that values are normally distributed
28
What does kurtosis mean?
The pointiness of the skew
29
What is the assumption of independent between subjects tests?
• Independent (between subjects) tests:
| o The assumption that distributions within treatment groups are normally distributed
30
What is the assumption of dependent (within subjects, paired) tests?
o The assumption that distributions of differences between treatment groups are normally distributed
31
What is the Shapiro-Wilk test in R?
o Shapiro.test
32
How can data be normalised?
o Every value in comparison needs to be treated equally and transformed
o Log () is used to bring data closer to the mean of values used for high values
o Sqr()- If some values are negative, a constant might have to be applied first so that all values become positive
o Reciprocal -> when very large values become very small
33
What is an outlier?
Any value >1.5x the interquartile range
34
What can you do to outliers?
Remove outliers
• mvoutlier
Transform data
35
What is the null hypothesis
Difference between means of treatment groups is zero
o Where H0 is the null hypothesis, which assumes that the difference between the observed value (data) and expected value (EV) is due to chance alone. p=p0 for example p=0.5
36
What is the alternative hypothesis?
• Alternative hypothesis- Difference between means of treatment groups is either not zero (two-tailed), or less/greater than zero (one-tailed). Can be given directionality
o Where H1 is the alternative hypothesis, which assumes that the difference between the observed value (data) and expected value (EV) is NOT due to chance alone. pp0 (upper sided test) or p doesn’t equal to p0 (two sided)
37
What is the assumption of homogeneity of variance?
All comparison groups have the same variance
38
How can the assumption of homogeneity of variance be tested for?
Levene’s test (leveneTest), p<0.05 indicates difference variance between treatment groups
39
What test does not assume homogeneity of variance?
• Welch’s t-test (t.test(paired=FALSE)) does not assume homogeneity of variance
o Need data to be normally distributed
40
When comparing 2 means between subjects, how can you make the data distributed using a test after transformation?
o Wilcox’s robust tests including (yuen)- trimmed mean calculated after 20% of scores have been removed from each extreme of the distribution
Can specify the level of trimming:
• Makes normal distribution
• If trim too much, data would lose its effect and a lot of experimental units would be needed
• The data should be transformed first
41
What 2 means have the smallest variance: between subjects or within subjects?
Within subjects
42
Why do we adjust value for plotting purposes?
Easier to see data trend
43
What test should we do if 2 samples have unequal variance?
Welch 2-sample T-test
| R command: t.test(x,y,mu=0,var.equal=F)
44
What test should we do if 2 samples suggest non-normality?
Transformations or non parametric tests
| Mann-Whitney-Wilcoxon test: wilcox.test
45
What test should we do if 2 samples are not independent
Paired T-Test
-Sometimes it is desirable to analyse dependent data. We often design an experiment to take advantage of this dependency in order to control variation between experimental groups
46
What is ANOVA?
o Analysis of variance (ANOVA) is used to compare several means between subjects
47
How do you see the ANOVA output?
o One-way independent ANOVA (aov(outcome~predictor)) then (summary) to see the output
48
What is assumed in ANOVA?
o Both homeogeneity of variance is assumed and normality is assumed
49
What is standard deviation?
the variance within one particular sample
50
What is standard error?
• Standard error- standard deviation of mean across many samples
51
What is relative quantitation?
compare abundance across multiple samples (typically expressed as a ratio) without determining number of molecules (e.g. copies, pmol/L, etc)
o Most commonly used because it’s easier
52
What is absolute quantitation?
etermine the number of molecules in each sample (e.g. copies, pmol/L, etc)
o Gives more information
o Don’t need to have multiple samples -> don’t need to compare between multiple sample (but generally need to compare against multiple samples)
o Can use known concentration of a thing, then use abundance of known thing to calculate exact molecules of an unknown thing
53
How does mass spectrometry work?
- Needs to be in a vacuum
1. You place the substance you want to study in a vacuum chamber inside the machine (into an inlet)
2. The substance is bombarded with a beam of electrons so the atoms or molecules it contains are turned into ions. This process is called ionization and produces molecular ions.
- Ion source
3. The ions shoot out from the vacuum chamber into a powerful electric field (the region that develops between two metal plates charged to high voltages), which makes them accelerate. Ions of different atoms have different amounts of electric charge, and the more highly charged ones are accelerated most, so the ions separate out according to the amount of charge they have. (This stage is a bit like the way electrons are accelerated inside an old-style, cathode-ray television.)
- Mass Analysers
4. The ion beam shoots into a magnetic field (the invisible, magnetically active region between the poles of a magnet). When moving particles with an electric charge enter a magnetic field, they bend into an arc, with lighter particles (and more positively charged ones) bending more than heavier ones (and more negatively charged ones). The ions split into a spectrum, with each different type of ion bent a different amount according to its mass and its electrical charge.
- Mass Analysers
5. A computerized, electrical detector records a spectrum pattern showing how many ions arrive for each mass/charge.
6. Computer
- Instrument control and data acquisition
54
What is the purpose of mass spectrometry?
• Can determine the structure and quantity of molecules by measuring their mass to charge ratio and comparing those to that of elements to see what elements are in the molecule.
55
Why is liquid chromotography useful?
o Use liquid chromatography to decrease complexity of sample before it’s fed into the mass spectrometer
56
How is liquid chromotography used?
o Separates out proteins/ peptides in time by introducing them more slowly to the instrument --> spread out separation of analytes in time so we can give mass spectrometer more time to quantify and identify number of analytes in the sample
Feed to the mass spectrometer more slowly
Instrument has finite speed at which it can operate
57
Where does liquid chromotography occur?
In the inlet of the mass spectrometer
58
Using liquid chromotography, in a 2 hour acquisition time, how many peaks would there be in a seperation and what does that mean?
o In a 2 hour acquisition (normal acquisition time), would have about 40,000 peaks in the separation- 40,000 analytes separated
59
What is electrospray ionization?
Electrospray ionization (ESI) is a technique used in mass spectrometry to produce ions using an electrospray in which a high voltage is applied to a liquid to create an aerosol. It is especially useful in producing ions from macromolecules because it overcomes the propensity of these molecules to fragment when ionized.
60
Why is electrospray ionisation done?
Applies very high voltage to liquid coming out of liquid chromatograph
Generates a spray of droplets which leads to generation of gas phase ions in front of the source region of the mass spectrometer -> ions get attracted and introduced into the machine itself.
61
What is tandem mass spectrometry?
o Technique to break down selected ions (precursor ions) into fragments (product ions)
o Fragments reveal aspects of the chemical structure of the precursor ions
62
What should the peaks of a mass spectrometer and liquid chromotography be?
Gaussian (normally distributed)
63
Analyse the peaks of mass spectrometers
• One large peak has many different peaks because all molecules analyzed are composed of different elements, which are made of many isotopes
o 1% of all carbon is C13, so that means that it is a common contaminant
o Very important for biological molecules (such as peptides and proteins)
o Rate of contamination by these isotopes are quite high
• Peaks are mixtures of different elements in the molecule being analysed
o Contamination becomes bigger and bigger as peaks decrease to the right of the monoisotopic peak
• These distributions are referred to as the isotopic envelope, and they represent the inclusion of isotopes within each particular analyte
64
What is a monoisotopic peak?
A pure peak
65
How does isotope-labelling for relative or absolute quantitation work in mass spectrometry?
• Quantitative mass spectrometry typically utilizes proteins labelled with heavy stable isotopes
• Labeled (heavy) peptides maintain the same characteristics as unlabeled or ‘light’ peptides and co-elute into the mass spectrometer from liquid chromatography columns
• In the mass spectrometer they are easily distinguished by their mass
• Algorithms are then used to extract the light and heavy peptide ion chromatograms, which represents the peptide’s abundance
• The light/heavy ratios are used to infer relative abundance
• By mixing the same labeled protein standard with different unlabeled protein samples, changes in relative abundance can be determined between biological conditions
For example;
if you applied a stimulus to the light medium and had the heavy medium as the control, then can identify which source it came from and hence identify the ratio difference between the two things.
66
How do you analyse proteins and metabolites using mass spectrometry (the entire pathway)
- Get a sample
- Homogenisation
- Protein extraction
- Protein quantitation
- Protein clean-up digestion
- Peptide clean-up
- LC-MS/MS analysis
- Data Processing (identify and quantify)
- Data analysis and vsualisation
67
How do you do the protein clean-up digestion step?
a. Simplify analytical technique
b. Digest in smaller components
c. Use proteases cut proteins up into smaller pieces in a very consistent way
d. Very sequence specific cutting
68
How does the LC-MS/MS analysis work?
a. LC- liquid chromatography
b. Tandem mass spectrometry look at what’s coming off chromatograph and choose to isolate individual analytes/peptides and determine what the masses of their fragments are work out structure of each analyte
c. Retention time- time of LC separation : want different analytes come off at different times
d. Map relative abundance indicated by colour
i. Sometimes, peaks are not Gaussian which causes problems
e. Can link identification data of masses of fragments to the abundance of the intact peptide
69
What is a TIC and what is its disadvantage?
• TIC- map of precursor ions coming off at any one time
| o Can’t use it for quantitative analysis
70
What is the information we can get from a TIC?
map of precursor ions coming off at any one time
o Can extract a single peak from data if we ask software to look specifically at abundance of one analyte
o Integrated area under the peak is proportional to the abundance of that analyte
One way of extracting piece of quantitative data
But still don’t know what that peak is
o Can isolate a peak, fragment it and see a new spectrum –Mass spectrum
Precursor peptide mass
Shows masses of fragments From gaps between fragment peaks, can determine sequence of peptide
71
What are the 5 options we can use to analyse proteins and metabolites?
- Abundance of intact precursor ion
- Abundance of peptide fragment ion
- Number of MS/MS per percursor
- Counting number of times you trigger a fragmentation spectrum on that particular analyte, because triggering is abundance-biased
- Extracting a three dimensional volume for each peak
72
```
How much variance can be introduced in:
-Metabolic labelling
-Label free
-Chemical labelling
Sort in order
```
Label free> chemical labelling > metabolic labelling
73
What is involved in metabolic labelling?
Feed enriched isotopically amino acids to cells
| o Enables us to introduce a light population of cells and a heavy population of cells
74
What does metabolic labelling enable us to do?
o Abundance difference between two peaks can be used for relative quantitation measuring
75
Why is metabolic labelling free of variation?
o Any bias in the workflow is applied to both heavy and light samples
76
How does insulin signalling pathway work?
• Insulin binds to insulin receptor
• Triggers phosphorylation cascades
o Phosphorylation of IRS1
o Recruitment of two subunits of PI3Kinase
P85 and p110
o Produce phospholipid called PIP3
o Recruits PDK1 and PDK/AKT
o AKT kinase acts on downstream substrates
o Ultimately culminates in movements of the vesicles
• Movement of glucose transporter from cytoplasm to plasma membrane
• Allows cell to take up glucose, especially in muscle cells
77
What are the keys to good imaging?
```
• Sample prep
o Extremely important
• Understand the technique
• Understand what makes a good image
• Reproducible unbiased analysis
```
78
What is an image?
```
• A visual representation of something
• Digital
• A 2d rectilinear array of pixels
o An xy table of values
• A way of capturing, storing and displaying data
• Transform an image using maths to extract data
• Made of pixels
o Pixel individual unit of an image
```
79
What are bits?
The basis of digital info (0 or 1)
80
What is bit depth?
The number of recorded bits per pixel
81
How many levels of intensity can n bit data store?
2^n
82
What is the benefit and cost of increased bit depth?
o Increased bit depth= increased information at the cost of file size
o Bit depth captures a lot more of the image
83
What is the trade off in imaging?
o Information capture per unit area vs file size
84
What is image resolution?
the number of pixels (wxh)
85
What is needed to make an image?
- A subject
- Energy source that will interact with the subject
- A way to control and focus the energy
- A detector
86
What is the common energy source in imaging?
EMR
87
What does an increase in wavelength mean?
A decrease in energy
88
What is the problem with imaging with UV, and what is a solution to this, as well as the problems to this solution?
o Imaging with UV causes damage to tissue and scatters more easily (doesn’t penetrate very far)
o Want to move into Infra-Red spectrum because deeper penetration and less damaging to tissues
Whilst there are benefits for going to IR, there’ll be loss of resolution or limit to resolution that can be achieved
89
How do you find Abbey's limit?
Wavelength/2
90
What is a way to control and focus energy, and what fields can you get as a result of this?
```
o Light changes in refractive index as it moves through the sample --> dramatically changes the info you can get from the image
o Results include
Brightfield
Phase contrast
DIC
Dark field
```
91
What are two types of inverted optical microscopes?
- Transmitted light path microscope
| - Reflected light path microscope
92
Why is the reflected light path microscope useful?
For samples such as metals or extremely thick organisms that remain opaque after ground that the transmitted light path microscope can't see- also mostly used for fluorescence
93
Why are inverted microscopes used and what is it?
* High magnification, high resolution, large working distance
* Typically used for observing cells on coverslips or surfaces close to coverslips submerged in liquid
An inverted microscope is a microscope with its light source and condenser on top, above the stage pointing down, while the objectives and turrets are below the stage pointing up.
94
When should brightfield be used?
Color-stained, high contrast sample
95
When should dark field be used?
Find structure, tiny sample
96
When should face contrast be used?
Low contrast, transparent sample
97
When should Differential Interference Contrast (DIC)?
Low contrast sample, for surface structure observation
98
What does widefield fluorescence microscopy do?
o Illuminates everything in the sample-> excites everything
o Illuminates the entire cell and captures info from entire cell -> image quality degraded as capture out of focus light
o Has low resolution
o All molecules out of plane of focus excited and makes the image noisy
o Epifluorescence- profound change in size but can’t see much change in intensity -> good for area
-Has 240 nm wavelength limit to resolution
99
What does confocal fluorescence microscopy do?
o Uses a pinhole near detector to cut out all out of focus light
o Get less resolution because 3 times resolution of light
o Cuts out of focus light
o Point scanning
o Spinning disk
Allows to look further into the cell
-Has 240 nm wavelength limit to resolution
100
Describe Total Internal Reflection Fluorescence and when you would use it
• Total Internal Reflection Fluorescence (TIRF)
o Light at critical angle gets completely reflected
o Generates weak ER wave that goes through interface and decays exponentially
o TIRF good for things that are close to the surface of the cell such as the plasma membrane
o Only allows image of membrane of cell
o Image of high resolution at basal membrane
-Has 240 nm wavelength resolution limit
101
Describe super resolution fluorescence microscopy
o Gives better resolution
o Stochastic methods- PALM, STORM, GSD
Statistics based
o Deconvolution methods- Zeiss Airyscan
102
What are types of fluorescence used for bioimaginging?
```
•Autofluorescence (label free)
o NAD and FAD- Redox
o Multiphoton harmonics-
2nd harmonics- collagen
3rd harmonics- RI mismatch; haemoglobin
• Antibody labelling
• Genetic encoding
o Fluorescent proteins
Genes of interest
Biosensors
o Dyes
Markers
Sensors
```
103
What are the different types of fluorescence microscopy?
- Widefield
- Confocal
- Total Internal Reflection Fluorescence
- Multiphoton
- Fluorescence lifetime imaging
- Super resolution
104
What is Glut4, where is it found and what can it be labelled with?
```
• Insulin responsive glucose transporter
• Necessary for insulin stimulated uptake of glucose
• Localization
o Perinuclear (around nucleus of a cell)
o Peripheral puncta
Partial overlap with endosomal markers
Specialized storage vesicles (GSVs)
o Translocate to the plasma membrane upon insulin stimulation
• Label with antibody
```
105
Once you have an image, what is the general workflow that preceeds it?
o Noise removal e.g. smoothing
Can come from camera or imaging techniques
Image histogram -> counts number of pixels at different intensity -> shows background noise that needs to be removed
o Background subtraction
o Segmentation
Thresholding
• Simple: gray scale images into binary images
• Works well on clean data
• Horrible for noisy data
o Make measurements
o Post analysis
106
How does machine learning work and what are its advantages/disadvantages?
```
o Pixel level
o Generates a set of descriptors
o Uses inferred background levels of noise to train the software
o Powerful approach
o Harder to implement
o Computationally expensive
o Also has downstream applications
```
107
Give a summary of GFP structure and where it comes from
• Beta barrel
• 11 beta strands
o Fluofor protected from outside environment by barrel
o If exposed to outside loses fluorescence
Made of Ser65- Tyr66- Gly674
Degrades into HBI
• Naturally occurring fluorescent protein from the jellyfish Aequorea Victoria
108
What are important properties of fluorescent proteins?
- Monomericity
- Brightness
- Spectral characteristics
- Sensitivity to pH
- Temperature sensitivity
- Lifetime
- Photobleaching
- Folding time
- Redox sensitivity
- Electrostatic potential
109
Do we want our fluorescent protein properties to be monomeric?
Yes
110
What is a side effect of brightness and light on the cell?
o The more light you have to put in, the more toxicity you get and the quicker you kill your cells
111
What is the excitation and emission wavelength of wild type GFP?
Excitation wavelength-488 nm peak which is blue
| Emission wavelength- 509 nm peak which is green
112
Why is temperature sensitivity important in fluorescent proteins?
Brightness will be affected by temperature
113
What is the lifetime of a protein?
o Probability of how long it is after photon is released after excitation
114
How does photobleaching impact the fluorescent proteins?
o Life cell imaging perspective-is important
o Excitation light forces the FP into a dark state- dead state- where it can no longer be excited and no longer releases energy
o Not good if you want to image for a long period of time
115
Why is knowing the folding time of a fluorescent protein important?
o Some will fold in minutes, others will take days
o Red FP go through phase where they’re green first
o Good as timers
o Time how a protein/ its abundance changes through time
116
What is a new fluorescent protein and what is it good for?
```
• 2016
• Small ultra-red fluorescent protein
o smURFP
• Developed into a new range of distinct fluorophores
• Require Biliverdin as a cofactor
• Ideal for in vivo imaging
```
117
What is GLUT4-eGFP used for?
o Cytoplasmic eGFP
o Intracellular trafficking
o Exocytosis-
Vesicle comes up to membrane and undergoes fusion
118
What are pHluorins and why are they used?
o A class of pH sensitive fluorescent proteins
Targeted mutation of eGFP
Theoretical pKa=7.11
o ~10 fold improvement in exocytotic signal
o Inside vesicle acidic compared to cytosol and outside of cell
o Not detected inside vesicle but as soon as released into cell it is detected (hence detects when vesicle undergoes fusion)
119
What are IRAP-pH and why are they used?
o Insulin Responsive Amino Peptidase
o Surrogate marker for GLUT4
o Trafics with GLUT4
~85% colocalization
o Lumenal pHluorin tag
o Insulin stimulates the fusion of IRAP-pH containing vesicles (GSVs)
o Insulin stimulates a transient burst of GSV fusion events
120
What are the components of pHluorin-GLUT4-TdTomato and what are they used for?
o Lumenal pHluorin (1st exofacial loop)
pH sensitivity
Marker of fusion
Exposed to external environment or inside the vesicle
o Cytoplasmic TdTomato
pKa 4.8
To see trafficking on the cytoplasmic side of the protein
o Can see both trafficking and fusion events
o rGLUTpHluor behaves like GLUT4
121
What is heterogeneity?
* Every cell appears to respond differently- heterogeneity
* Important component of biology
* Seen in single cells and complex organisms
* Seen as an outcome of randomness but results can be replicated, so this is not a sufficient explanation
* Heterogeneity is present with many doses of insulin
* Akt is driving the heterogeneity in GLUT4
122
What is Akt?
* Fluorescent protein at N terminus
| * eGFP Akt2 is the classic probe for this
123
What is the difference between eGFP and TagRFT-T Akt and why is this?
• Green (eGFP) and red (TagRFP-T) Akt behave differently
o This is because of their net charges:
eGFP is has a net charge of -5
tag RFP-T has a net charge of 0
o Due to its negative charge, eGFP can’t go to membrane properly so less efficient in showing what’s going on there than tag RFP-T
• Electrostatic potential can alter fusion protein behaviour
124
How do you read files in R?
• Read files such as data.txt using read.delim, scan or read.table
125
What can a vector in R be?
Numeric
Character
Logical
126
What does a function do in R?
• A function returns a value resulting from programming statements. A function may or may not have input arguments.
127
How do you make a matrix?
o Mymatrix
128
In R, if you wanted the first row and the second column number of your matrix, what would you do?
Mymatrix[1,2]
| • Want the first row second column number
129
What is a population?
o The set of data (numeric or otherwise) corresponding to the entire collection of units about which information is sought
130
What is a sample?
A subset of the population data that are actually collected in the course of a study
131
Why do we need a sample?
o In most studies, it is difficult to obtain information about the whole population which is why we rely on samples to make estimates and inferences related to the whole population
132
What is the central dogma of statistics?
Population -> Probability -> Sample -> Inference
133
What is the difference between a parameter and a statistic?
Parameter:
Number that describes a population
Denoted in greek letters
Unknown fixed number
Statistic:
Number that describes a sample
Denoted in roman letters
Variable whose value varies from sample to sample
134
What are the mean, median and mode measures of?
Location
135
What is the sample mean?
The sum of all the observations divided by the number of observations
136
What is the median?
median of a set of data is a value such that at least one half of the observations are less than or equal to x and at least one half of the observations are greater than or equal to x
o Sample median- the (n+1)/2th largest observation if n is odd
o The average of the (n/2 +1)th largest observation if n is even
• The median is not influenced as much as the mean by outliers because it is robust
137
What is the mode?
the most frequently occurring value among all observations in a sample
o If no entry is repeated, the data set has no mode
o If two entries occur with the same greatest frequency, each entry is a mode
138
Compare the mean and median for symmetric and skewed data
* For symmetric data, the mean is usually less variable from sample to sample than the median
* For skewed data, the median is a better measure of location
139
What are the 3 measures of spread?
Standard deviation, MAD (median absolute deviation), IQR
140
What is the range of a list, and what is the problem with this?
• Range of a list is the largest value minus the smallest value
o Misleading because it is solely influenced by two most extreme values
141
What is deviation?
the difference between the individual sample points and the average
142
How do you find Median Absolute Deviation (MAD)?
MAD= median(| Xi- median(X)|)
143
What does standard deviation (variance) do and how do you find it?
o Find the mean of the data
o Make a list of deviations from the mean (value-mean)
o Calculate the average of the squares of deviations (var)
144
What is the IQR range and what is an advantage of it?
```
• The three quartiles, Q1, Q2 and Q3 approximately divide as a ordered data set into four equal parts
o The interquartile range is defined as the upper quartile (75th percentile) minus the lower quartile (25th percentile). Contains middle 50% of data
o IQR is robust
o quantile(x)
```
145
How do you do a grouped bar plot in R?
o counts
146
How do you do a boxplot in R?
boxplot(mpg~cyl,data=mtcars, main="Car Milage Data", xlab="Number of Cylinders", ylab="Miles Per Gallon")
147
How do you do a histogram in R?
hist(mtcars$mpg, breaks=12, col="red")
148
What is Tableau?
Tableau is a database that allows visualization across data tables
149
What is the identifier column in tableau?
• Identifier column- primary key column that allows linkage of one table to another
o Name of column doesn’t have to be identical but the values do
o But make sure you’re not cutting a lot of data when you’re putting the tables together
150
What is the join in tableau?
links columns from 2 or more tables in a relational database. Linked via a primary key
151
What is the primary key in tableau?
a minimal set of columns that uniquely specify a row in a table
152
What are the 4 different types of joins?
Inner- any entry only found in one table is discarded
Left- keep all entries in one side of the join linked to the other
Right- keep all entries in one side of the join linked to the other
Full outer
153
What is a union and the requirements for a union?
• Union- combines table by stacking with the same number of columns and compatible data types
o Requires same amount of columns and data types
o Use it for repetitive experiments
154
What are dimensions in tableau?
contain qualitative values. You can use dimensions to categorize, segment and reveal the details in your data
155
What are measures in tableau?
contain quantitative values that can be described numerically and then either displayed as they are, aggregated, or used for mathematical operations
156
What is aggregation in tableau?
summarise multiple data points (e.g. mean, median, sum…)
157
How do you create a new column in Tableau?
dimension 1+ “separator” + dimension 2
158
What are IF statements in tableau?
o IF CONTAINS( criteria, “criteria”) =TRUE
o THEN “criteria”
o ELSE “Something else”
o END
159
What is quantitative in tableau?
Decimal measurements
160
What is ordinal in tableau?
Follows a sequence
161
What is nominal in tableau?
Naming conventions
162
What level of interactivity do we have in tableau?
• Sheets can be integrated in web pages
• Tootips pop-up data
• Can drag measures and dimensions in an already made plot and the plot will change
• Discrete filters
o Plot will update depending on what filters you tick on it
• Continuous filters
o Slide bar
• Can make word clouds
o Size of word can be proportional to amount of times word is in table
