Research Exam Semester 2 – 2019-2020 June 12, 2020
Download this file and add your answers. If it is more convenient (as for sketches and equations) you can take pictures of hand-written pages and insert those into the docu- ment.
Don't forget to save regularly!
Submit your finished document before 11:30 via Brightspace: 1920 Research Biomedical Science > Activities > Assignments > Upload Research Exam Semester 2
(Students with special facilities can submit until 12:00, and have a separate submission module)
In case of technical problems, mail T.Oostendorp@donders.ru.nl or call 06 81 89 41 20.
The questions must be answered in English. If you cannot remember a specific English term, you may use the Dutch term.
Be precise in your answers. Adding correct but irrelevant information will not increase your score. Adding incorrect information, even if it is irrelevant, will lower your score.
During the exam, you may want to consult these books (see link in Brightspace):
• Baynes & Dominiczak: Medical Biochemistry
• Campbell: Statistics at square one
• Donders: Literature Measurement errors
• Fletcher: Clinical Epidemiology
• van Oosterom en Oostendorp: Medische Fysica
• Petrie and Sabin: Medical Statistics at a Glance
• Turnpenny: Emery's Elements of Medical Genetics
• Form with statistical formula’s
Name:
Student Number:
Question 1
Q3: What we can learn from urine – dr. H. Pluk (20 points)
Urinary sucrose and fructose to validate self-reported sugar intake in children and adolescents: results from the I.Family study
Intemann et al. European Journal of Nutrition volume 58, pages1247–1258 (2019) Purpose
Excessive consumption of free sugar increases the risk for non-communicable diseases where a proper assessment of this intake is necessary to correctly estimate its association with certain diseases. Urinary sugars have been suggested as objective biomarkers for total and free sugar intake in adults but less is known about this marker in children and adolescents. Therefore, the aim of this study is to evaluate the relative validity of self- reported sugar intake using urinary sugars in children and adolescents.
Methods
The study was conducted in a sample of 228 participants aged 5–18 years of the I.Family study that investigates the determinants of food choices, lifestyle and health in European families. Total, free and intrinsic sugar intake (g/day) and sugar density (g/1000 kcal) were assessed using 24-h dietary recalls (24HDRs). At least 19 children and adolescents per center, of both sexes, agreed to recall their diet using a 24HDR and to provide a morning urine sample on the same day of the dietary recall. Urinary sucrose (USUC) and urinary fructose (UFRU) were measured in the urine samples. Correlation coefficients and regression models were used to investigate the relationship between intake of different types of sugar and urinary sugars.
Results
The Pearson correlation between usual sugar density calculated from multiple 24HDRs and the sum of USUC/Cr and UFRU/Cr (USUC/Cr + UFRU/Cr) was 0.40. Linear
regression models showed statistically significant positive associations between USUC/Cr + UFRU/Cr and the intake of total and free sugar.
Conclusions
These findings support the relative validity of total and free sugar intake assessed by self- reported 24HDRs in children and adolescents.
A. The assessment of dietary intake (by self-reported 24-h dietary recalls) is especially challenging in children. Give three reasons why. (3 points)
Answers (1 point each correct reason):
- because children have a highly variable diet, which makes it more difficult to recall.
- recall methods such as the 24-h dietary recall (24HDR) highly depend on the children’s cognitive abilities to remember and to correctly estimate food quantities
=> is more difficult for them than for adults.
- If proxies report for the children 24HDRs will depend on the proxies’ memory and their presence during children’s meals => extra step involved
- Children are less able to read and write than adults
The methods of the presented study were described as follows:
Laboratory methods
Collected morning urine samples from the eight I.Family study centers (located in Belgium, Cyprus, Estonia, Germany, Hungary, Italy, Spain, and Sweden) were sent for analysis to the Dept. of Biochemistry of the University of Naples under standard shipping conditions (either at − 20 °C or at − 80 °C). Urinary sucrose (USUC) and urinary fructose (UFRU) concentrations were determined using an enzyme-based kit (sucrose/D-glucose/D- fructose from Boehringer Mannheim) and a Perkin Elmer Lambda Array
spectrophotometer to measure the absorbance rate. All determinations were run in triplicate. Detected concentrations were in the range of 1–150 mg/L. Within this detection range, linearity of measurements was observed. Assay control solutions in the range of expected values for sucrose and fructose were provided in the enzyme-based kit. Values for this control were remarkably stable (glucose concentration: 100.6 ± 5.9 mg/L,
mean ± standard deviation, coefficient of variation 5.9%).
Analyses
Urinary concentrations of sucrose and fructose for all participants were expressed as mg/L of urine and also as mg/g of creatinine (USUC/Cr, UFRU/Cr). The (mean/usual) sugar density was calculated as (mean/usual) total sugar intake (in gram) per 1000 kcal of (mean/usual) total energy intake.
B. Identify five critical steps in the experimental setup (as presented in the abstract and methods section above); explain shortly why these are critical steps to get valid results in this experiment. (10 points)
Answers (2 each correct answer):
- Morning urine sample was provided on the same day of the dietary recall => for proper alignment/mechanistic explanation of biomarker the urine sample should be properly related to timing of recall.
- All samples from different centers were sent for analysis to the Dept. of
Biochemistry => all analysis done by one laboratory, less spread in measurement errors.
- Urine shipping conditions were either at − 20 °C or at − 80 °C => cold temperature to preserve samples and prevent breakdown of your biomarker.
- All determinations were run in triplicate => to check for measurement errors.
- Validated kits and/or machines were used.
- Linearity of measurements was observed in detection range => otherwise you cannot compare samples over a range of values.
- Assay control solutions were used, in the same range as the expected values with very stable values => control sample to check for standard errors during procedure.
C. Explain why urinary concentrations of sucrose and fructose were expressed as mg/L of urine and also as mg/g of creatinine (USUC/Cr, UFRU/Cr). (3 points)
Answer: To correct for fluctuations in urine volume (urine dilution), since these are spot sample urines and not 24 hrs urine samples. If you use mg/L you cannot compensate for urine dilution in spot samples (you do not know the total volume) and you may be less accurate.
Note: creatinine secretion depend on muscle mass which is still built up in children
=> complicating factor. GFR is not a correct answer => 0 pts
The researchers state the following result: “The Pearson correlation between usual sugar density calculated from multiple 24HDRs and the sum of USUC/Cr and UFRU/Cr
(USUC/Cr + UFRU/Cr) was 0.40. Linear regression models showed statistically significant positive associations between USUC/Cr + UFRU/Cr and the intake of total and free sugar.”
D. Sketch a graph that represents this Pearson correlation. Select the best fitting representation of this correlation from the examples below (circle the graph you choose, or mention the number of the representation (1 to 14)) and complete the graph by naming the x- and y-axis. If you cannot sketch explain in words what you would see in the intended graph and name the axes. (4 points)
1 2 3 4 5 6 7
8 9 10 11 12 13 14
Answer:
Scatterplot with on x-axis: usual sugar density (g/1000 kcal) and on y-axis:
USUC/Cr + UFRU/Cr (mg/g)
1 point for correct labelling x-as => usual sugar density (g/1000 kcal) 1 point for correct labelling y-as => USUC/Cr + UFRU/Cr (mg/g)
2 points for medium correlation Pearson correlation coefficient 0.40 => cloud of dots with a positive slope. Abstract: 228 participants => few single points can be
concerned.
Best fitting:
No values given for (USUC/Cr + UFRU/Cr) and sugar intake => no numbers on axes possible. Units can be given (see methods: “Concentrations of sucrose and fructose for all participants were expressed as mg/L of urine and also as mg/g of creatinine (USUC/Cr, UFRU/Cr)” => Unit mg/g
“The (mean/usual) sugar density was calculated as (mean/usual) total sugar intake (in gram) per 1000 kcal of (mean/usual) total energy intake” => unit g/1000 kcal.
Name:
Student Number:
Question 2
Q4: Modelling Epidemic Outbreaks – dr. T. Oostendorp (15 points)
The standard model that is used to predict the number of infectious people in an epidemic is the SIR model. In this model, the differential equations for the number of susceptible people 𝑆(𝑡) and infectious people 𝐼(𝑡) are
𝑑
𝑑𝑡𝑆(𝑡) = −𝑅!
𝐷 𝐼(𝑡) 𝑆(𝑡) 𝑁 𝑑
𝑑𝑡𝐼(𝑡) =𝑅!
𝐷 𝐼(𝑡) 𝑆(𝑡)
𝑁 − 𝐼(𝑡) 𝐷 with
𝑁 population size
𝑅! basic reproductive number
𝐷 average duration of infectiveness in days
A. Explain why the minus term in the differential equation of 𝐼(𝑡) is −𝐼(𝑡)/𝐷. (4 pts)
Change = in - out
The minus term represents "out": the number of infectious people per day that stop being infectious.
On average, people are infective for 𝐷 days. Consequently, every day the fraction 1/𝐷 of the infectious people stop being infectious. Hence the total number of people that stop be- ing infectious is 1/𝐷 ∙ 𝐼(𝑡)
Note: there are two versions of question B; the first one is for students that did the module Modelling Epidemic Outbreaks in 2020 and learn R, the second one is for those that did it in the years before and learned Simulink.
Version 1 (students in 2020; R):
B. Below is a skeleton for a program in R that implements the SIR-model. Replace the three dotted lines by the code needed in order to make the script compute the number of infected people. (5 pts)
simDuration <- 100 deltaT <- 1
nSteps <- simDuration/deltaT N <- 100
R0 <- 2 D <- 7
S <- numeric(nSteps) I <- numeric(nSteps) T[1] <- 0
I[1] <- 0 S[1] <- N-I[1]
for (i in 1:Steps) {
...
newResistants <- I[i]/D * deltaT T[i+1] <- T[i] + deltaT
...
...
}
plot(T, I, type='l', xlab='t (days)', ylab='I(t)')
newInfections <- (R0/D) * (S[i]/N) * I[i] * deltaT S[i+1] <- S[i] - newInfections
I[i+1] <- I[i] + newInfections – newResistants note: this box contains the answer and must be removed in the final version
Version 2 (earlier students; Simulink):
B. Below is an unfinished Simulink diagram that implements the SIR model. What type of block should be inserted at the red cross, and what name should be given to that block? (5 pts)
An integrator block, it should be named I(t).
Let’s assume that in 2025, the covid-25 pandemic will develop. Social distancing will again be enforced in order to bring down the value of 𝑅! for covid-25. The RIVM (Dutch Institute for Public Health and environment) states that the number of infectious people will de- crease if that value can be kept below 1.
C. Explain why the number of infectious people will decrease when 𝑅! < 1. (3 pts) If 𝑅! < 1, each infectious individual will infect – on average – less than 1 person, even if everybody else is susceptible. As a result, each new ‘generation’ of infectious people will be smaller.
The plot in figure 2.1 shows the result of the SIR model for corona, where at day 40 the value of 𝑅! changed instantaneously from 2 to 1.3.
Figure 2.1 Predicted number of infectious people if 𝑅! changes from 2 to 1.3 on day 40.
D. Explain why the number of infectious people decreases after 40 days, even though at that moment 𝑅!>1. (3 pts)
As the epidemic has already been going on for at least 40 days, not everybody can be susceptible anymore. Consequently, each infectious individual will infect only 𝑅!"($)
& peo-
ple. Apparently, 𝑆(𝑡) has dropped enough for this value to be below 1.
0 20 40 60 80 100
0246810
t (days)
I(t) (%)
Name:
Student Number:
Question 3
Q4: Population research: Associations and causal relations – dr. F. de Vegt (20 points)
Use ‘Elnaz Daneshzad et al - Association between a low-carbohydrate diet and sleep status, depression, anxiety, and stress score – abstract and tables.’
A. What is the research question in the study of Daneshzad et al?
In addition, specify the determinant, the outcome and the study population (3 pts).
Research question: What is the association between a low-carbohydrate diet (LCD) and sleep and mental status (depression, anxiety and stress) in women with type 2 dia- betes mellitus?
Determinant: carbohydrate consumption, low-carbohydrate diet Outcome: sleep status, depression, anxiety and stress
Study population: 265 women with type 2 diabetes mellitus, (who referred to health centers or diabetes research in Tehran, Iran)
B. The study design was a cross-sectional study. Describe the set-up of a cross-sectional study and mention one advantage and one disadvantage of this study design (3 pts).
A cross-sectional study takes place at a single point in time. It does not involve manipu- lating variables. Data collection is at one single point in time, which allows researchers to look at numerous characteristics in a study population at once (+) at low costs (+).
However, as this is a snapshot, there is no time between exposure and outcome, and causality cannot be assessed (-)
C. Instead of a cross-sectional study, researchers could have designed a randomised controlled trial (RCT) to investigate the same research question. How does a RCT look like for this research question? (3 pts)
Answer – Assemble a cohort of women with diabetes mellitus type 2. Use a randomisa- tion procedure to make two comparable groups of women. One group will follow a low- carbohydrate diet for three months; the other group can maintain their normal dietary habits (control group). Measure dietary intake (carbohydrates!) at the start and each month in the trial. Measure also sleep status and mental status at baseline, month 1, 2 and after the trial. All measurements should be done in a blinded way.
D. See Table 4 in the abstracts and tables. In total 188 subjects suffer from poor sleep. If per quartile of LCD score the numbers of patients were given, which test must be used to determine whether the percentage ‘poor sleepers’ differs among the quartiles? (1 pt) Chi-square test.
E. In Table 4 four outcome parameters are analysed with binary logistic regression. The statistical section mentioned ‘…and the significance value set as P < 0.05’.
Which methodological problems did the authors not deal with? (3 pts)
By using a lot of statistical tests, the overall probability of finding a false significant re- sult is inflated. Some correction for multiple testing should be applied or mentioned in the discussion.
F. What is the interpretation of the result 0.44 (0.19; 0.98) mentioned in Table 4? (3 pts) Women in quartile 4 of the LCD score (which corresponds to the lowest CH consump- tion!) have a 56% / 0.44 times lower risk for stress compared to the women in quartile 1 of the LCD score (meaning the highest CH consumption). As the value of 1 is not in the 95% confidence interval, this result is statistically significant. The model used was ad- justed for age, BMI, energy intake, SES, nap time, night sleep, physical activity, medi- cation, and supplement consumption.
G. Explain why ‘duration of diabetes’ might be a confounder in the research of Daneshzad et al. (2 pts)
‘Duration of diabetes’ in itself may be a risk factor for sleep problems and mental sta- tus; having a disease for a longer time may be stressful. There may also be an associ- ation between ‘duration of diabetes’ and carbohydrate intake, as patients may have changed and adapted their diet to their disease status. Therefore, the main research question (association between LCD and sleep + mental status) may be confounded by age.
H. What has been done in the research of Daneshzad et al. to control for confounders?
(2 pts)
A multivariate regression model has been used to adjust for confounding.
Name:
Student Number:
Question 4
Q4: T- and B- cells in the lab – dr. E. Blaney Davidson (20 points)
Amanda is researching T-cells in the spleen. She wants to visualize the T-cell receptor (TCR) using immunohistochemistry (IHC). Amanda made formalin fixed sections from hu- man spleen on glass and stained these for the T-cell receptor TCR, but found no staining.
A. Name three experimental factors related to the IHC protocol that could have caused the lack of staining. (3 pts)
(3 points: one for each item)
• The primary antibody and the secondary antibody are not compatible
• The primary antibody was forgotten
• The primary/secondary antibody has become inactive (incorrect storage)
• Antibody concentration was too low
• Antibody not suitable for IHC procedure or parafine
• Tissue material too old
• The protein is not present in the tissue of interest
• Deparaffinization may be insufficient
• Antigen retrieval may be insufficient
• Incubation time was too short
• In case of fluorescence microscopy, the secondary antibody was not stored in the dark
• In case of fluorescence microscopy, the sample was left too long in the light after IHC procedure
• In case of fluorescence microscopy, use of the wrong mounting medium
B. After having figured out what went wrong in 1a, Amanda repeats the immunohisto- chemistry for TCR of the spleen. But while incubating the 3,3ʹ-Diaminobenzidine on her sections, she is interrupted by an important phone call and forgets about her staining for a while. When she finishes her phone call and gets back she is doubtful about her staining. Can she still objectively use her immunohistochemically stained sections to interpret the amount of TCR in the spleen? Explain why she can or cannot. (3pts)
She cannot. At this step HRP reacts with H2O2 and DAB (the chromogen) to form a brown colored deposit and H2O. Importantly, HRP (an enzyme) can react again and again with a substrate molecule to create more reaction product. This is why it is best to keep monitoring what happens on the slides. When the reaction is not stopped a lot of reaction product is produced and the slides will be super brown. Therefore, you won’t be able to state anything about amounts.
A patient recovered of Covid-19 displays antibodies to COVID-19 in the circulation. An in- direct ELISA is a suitable method to measure antibody levels in the circulation. An indirect ELISA means that you have a combination of a detecting antibodies and a labeled conju- gate.
The following materials are available in the lab: standard ELISA plates, blocking buffers, washing buffers, substrate for horse radish peroxidase (HRP) and an ELISA plate reader.
The following antibodies, plasma and antigens are available in the lab:
• Avidin-HRP conjugate
• Guinea pig(Ig) anti-rat Ig
• Biotinylated Hamster(Ig) anti-rat Ig
• Rabbit(Ig) anti-human Ig
• Rat (Ig) anti-human Ig
• Biotinylated Rat(Ig) anti-rabbit Ig
• COVID-19 spike protein solution
• Plasma of a patient who recovered from Covid-19.
C. Write down the ELISA procedure to determine the level of anti-COVID-19 antibodies in blood of a recovered patient in a step-wise manner. In total there are 6 steps. Only indi- cate the sequence of steps regarding the application of antigen and antibodies that are required in this indirect ELISA (so blocking and washing steps do not need to be indi- cated).(4 pts)
Answer:
1. Coat plates with COVID-19 spike protein 2. Incubate with plasma of recovered COVID pts 3. Add Rabbit anti-human Ig
4. Add Biotinylated Rat(Ig) anti-rabbit Ig 5. Add Avidin-HRP conjugate
6. Add HRP substrate
(All steps in correct sequence is 4 points; Steps mentioned, but not in correct order, each 0.5 points)
D. What is the reason to include a blocking step?
Answer:
To prevent non-specific binding of antibody (1 pt)
In order to determine whether a change in cytokine concentration has influenced the com- position of lymphocyte subpopulations in blood, you like to use flow cytometry to investi- gate the four main lymphocyte populations (Helper T cells, Cytotoxic T cells, B cells and NK cells). You use a blood sample for your analysis.
E. Which antibody combinations should be used in flow cytometry to determine each pop- ulation of lymphocytes in the blood separately? Provide an answer for the four main lymphocyte populations. (4 points)
• Helper T cells: CD45 and CD3 in combination with CD4
• Cytotoxic T cells: CD45 and CD3 in combination with CD8
• B-cells: CD45 and CD19
• NK cells: CD45 and CD3 in combination with CD56 (1 point per correct item)
F. How should you perform the flow cytometry analysis to evaluate correctly the lympho- cyte populations? Describe the different plots by mentioning their axes in the correct rank order, starting with forward scatter/ side scatter. Describe which cell population(s) can be selected in each plot by using this rank order. (5 points)
• 1. FS/SS (Selection: all white blood cells and exclusion of debris) – 1pt
• 2. CD45/SS (Selection: to gate on lymphocytes within all white blood cells), - 1pt
• 3. Within lymphocytes the populations will be selected by:
o Plot a.: helper and cytotoxic T cells (CD3+/CD45+ followed by
CD4+CD8+). (Selection of resp. Helper T cells and Cytotoxic T cells) – 1pt
o Plot b.: B cells (CD19+/CD45+) (Selection of B cells) – 1pt o Plot c.: NK cells (CD3-CD56+) (Selection of NK cells) – 1pt
