A2.4

Question 1 Single best answer 1 point

A gene is differentially expressed between drug-treated and vehicle-treated mouse liver. What does that statement say about the gene?

  1. Its expected expression changes at least twofold, up or down, between the two conditions
  2. Its counts differ between the particular treated and vehicle samples that were sequenced
  3. Its expected share of the RNA in a sample differs between conditions
  4. Its expected number of RNA molecules per cell, which the read counts measure, differs between conditions

Question 2 Single best answer 1 point

An analysis of drought-stressed against well-watered Arabidopsis leaves reports two separate answers for each gene. What are they?

  1. Evidence that its expression changed, and the size of that change
  2. Whether the gene is expressed at all, and how much it is expressed
  3. Its expression in drought leaves, and its expression in watered leaves
  4. The p-value of the change, and the adjusted p-value of the change

Question 3 Single best answer 1 point

E. coli cultures exposed to an antibiotic give these \(\log_2\) fold changes: recA \(+3\), acrB \(-5\), lacZ \(+1\), ompF \(-2\), marA \(+4\). Which order lists the genes from the largest change to the smallest?

  1. lacZ, ompF, recA, marA, acrB
  2. acrB, marA, recA, ompF, lacZ
  3. marA, recA, lacZ, ompF, acrB
  4. marA, recA, lacZ, acrB, ompF

Question 4 Single best answer 1 point

Across four replicate Drosophila wing discs, vg has mean count \(400\) and variance \(400\), and wg has mean count \(20\) and variance \(180\). Which genes show overdispersion?

  1. Both genes
  2. Neither gene
  3. Only vg
  4. Only wg

Question 5 Single best answer 1 point

Three pesticide-exposed and three unexposed honeybee hives are compared. The ten bees of each hive are pooled, and each pool is sequenced in two lanes. Which repeats are biological replicates, and how many per condition?

  1. Two; the lanes are biological and the hives are technical
  2. Three; the hives are biological and the lanes are technical
  3. Thirty; the pooled bees are biological and the lanes are technical
  4. Six; the lanes are biological, since each lane is its own column of counts

Question 6 Single best answer 1 point

A tomato experiment grows one plant under each of two treatments and sequences one RNA sample from each plant. Why can this design not support a conclusion that a gene changed?

  1. Sequencing samples reads at random, so one count per plant is too noisy to compare
  2. Two samples are too few to put on a common scale, so their counts are not comparable
  3. Each RNA sample was sequenced only once, and a repeat run of each would supply the missing variation
  4. The variation between plants treated alike is unknown, so the difference has nothing to be judged against

Question 7 Single best answer 1 point

A test of one gene compares zebrafish fins from knockdown and control animals. What null hypothesis does the test use for that gene?

  1. Its true \(\log_2\) fold change between the two conditions is exactly zero
  2. Its observed \(\log_2\) fold change in the sequenced samples is exactly zero
  3. Its counts agree closely across the replicates of each condition
  4. Its p-value is above 0.05 when the two conditions are compared

Question 8 Single best answer 1 point

A Daphnia pesticide comparison gives one gene a p-value of \(0.03\). What does that p-value say?

  1. If the gene did change, about 3 in 100 repeats would give a difference smaller than this one
  2. If the gene did not change, about 3 in 100 repeats would give a difference at least as large
  3. The gene is expressed about 3 percent differently in the pesticide and control animals
  4. If the experiment were repeated, about 97 in 100 repeats would give a difference at least as large

Question 9 Single best answer 1 point

A rice gene under salt stress has \(p = 0.001\). A student says there is a 0.1% chance that the gene is unchanged. What is wrong with that statement?

  1. The p-value is estimated from a few replicates, so it is only a rough measure of that chance
  2. The p-value is worked out by supposing the gene is unchanged, so it does not measure the chance that it is unchanged
  3. The p-value is calculated for one gene, so the adjusted value, which corrects for the other genes tested, is that chance
  4. The p-value is the chance that the result would repeat in a second experiment, not that the gene is unchanged

Question 10 Single best answer 1 point

Two groups of identically treated C. elegans are compared gene by gene, and hundreds of genes fall below \(p = 0.05\). Why does this happen?

  1. About 5% of genes truly change between two groups of worms, and thousands of genes were tested
  2. Testing thousands of genes makes the cutoff easier for each gene to pass, so more of them fall below it
  3. A gene whose counts differ between the groups falls below \(p = 0.05\), and thousands of genes differ a little
  4. Each unchanged gene falls below \(p = 0.05\) about one time in twenty, and thousands of genes were tested

Question 11 Single best answer 1 point

A pig skeletal muscle comparison in which no gene changed is tested at p < 0.05 on two platforms, one covering 4,000 genes and the other 16,000. About how many more genes pass on the larger platform?

  1. 12,000
  2. 0
  3. 800
  4. 600

Question 12 Single best answer 1 point

A chicken embryo limb bud comparison calls a list of 400 genes at a 10% false discovery rate. What does that rate say about the list?

  1. No more than about 10% of the genes tested, called or not, are expected to be unchanged, though the analysis does not say which ones
  2. No more than about 360 of the 400 genes are expected to be unchanged, since 10% is the share that changed
  3. No more than about 40 of the 400 genes are expected to be unchanged, but the analysis cannot say which ones
  4. Each of the 400 genes has at most about a 10% chance of being unchanged, the same for each gene

Question 13 Single best answer 1 point

In heat-shocked against control yeast, ACT1 has \(\log_2\) fold change \(0.1\) and adjusted \(p = 10^{-6}\), and SPS100 has \(\log_2\) fold change \(3\) and adjusted \(p = 0.6\). How is each gene best classified?

  1. ACT1: small change, weak evidence; SPS100: large change, strong evidence
  2. ACT1: small change, strong evidence; SPS100: large change, weak evidence
  3. ACT1: small change, strong evidence; SPS100: large change, strong evidence
  4. ACT1: small change, weak evidence; SPS100: large change, weak evidence

Question 14 Single best answer 1 point

Nitrogen-starved against control maize roots give these \(\log_2\) fold changes and adjusted p-values. nrt1: \(3\) and \(0.001\); nir1: \(-4\) and \(0.01\); amt1: \(5\) and \(0.4\); glt1: \(0.5\) and \(0.0001\); gln1: \(-3\) and \(0.2\); nrt2: \(4\) and \(0.03\). Which genes are on the hit list with cutoffs of 2 on the absolute \(\log_2\) fold change and 0.05 on the adjusted p-value?

  1. nrt1, nir1, glt1, nrt2
  2. nrt1, nir1, amt1, gln1, nrt2
  3. nrt1, nrt2
  4. nrt1, nir1, nrt2

Question 15 Single best answer 1 point

Sporulating against growing Bacillus cells give a hit list at 1 on \(\log_2\) fold change and 0.05 on adjusted \(p\). A collaborator filters the same table at 2 and 0.01. How does her list compare, and what do its adjusted \(p\)-values test against?

  1. It is the same size or shorter, and each adjusted \(p\) tests against a change of 2 or less
  2. It is the same size or shorter, and each adjusted \(p\) tests against no change
  3. It is the same size or longer, and each adjusted \(p\) tests against a change of 2 or less
  4. It is the same size or longer, and each adjusted \(p\) tests against no change
Back to top