Read the summary and the most important questions on Biosystems Data Analysis

• A: Data properties, preprocessing, p-value

  This is a preview. There are 26 more flashcards available for chapter 04/01/2021
  Show more cards here

• OMICS is not high-troughput data. Explain what it is then?

  • High troughput = quantification of  a single component in a large number of samples (in short time)
  • so determine one concentration in many samples
  • Multiplex technologies = quantification of a large number of (related) components in a single sample
  • --> OMICS

• Give an overview of the RNA-seq procedure.

  • Stop all activity in the sample (quenching)
    
  • Isolate mRNA
  • Reverse transcription (RNA --> DNA)
  • Optional amplification by PCR
  • Library construction (attaching sequence tags/adaptors)
  • Sequencing

• Which points should you consider in regard to quenching?

  • RNA's have short half-lifes in cells
    
  • RNAases have to be stopped
  • Stress of handling can induce gene expression
  • Breaking cells (bacteria) can be hard
  • Obtaining sample can be time-consuming

• Which points should you consider for RNA isolation?

  • Most RNA is ribosomal RNA
    
  • Eukaryotic mRNA can be enriched using poly-A hybridization

• Name the two types of variation.

  • Biological variation
    
  • = variation of interest
  • Variation between similar samples (individuals)
  • Technical variation
  • filter out

• Name sources of technical variation (eg RNAseq).

  • Sample preparation (media, temperature...)
    
  • Sample isolation (handling, speed of sequencing)
  • Differences in mRNA quality
  • cDNA synthesis
  • Amount of cDNA added
  • Sequence bias
  • Random measurement error (you cannot get rid of it unless you repeat it many times and take the average)

• Where does bias in RNA seq come from?

  • Variation in amount of isolated mRNA
    
  • Variation in quality of isolated mRNA
    
  • Variation in quenching efficiency
  • Variation in cDNA synthesis efficiency
  • Variation in sequencing efficiency (number of sequences read)

  
  Or, due to interesting biological variation in amount of mRNA!

• Give 3 subtle bias effects in RNA-seq.

  • Fragment length (size selection)
    
  • Position (degredation)
  • Sequence bias (high GC --> lower counts)

• When executing normalization, you need to assume hypotheses about the origin of observed variations in sequencing counts. Give the two hypotheses and subsequent procedures.

  • H1: Approximately equal concentration of mRNA in each sample
  • Implies variations in total counts per sample are due to technical reasons
  • Solution: divide sequence count for each gene by the total sequence count of the sample
  • Result: RPM: Rate Per Million reads
  • H2: Approximately constant number of sequences per kilobase of mRNA
  • Variations in counts between genes are due to gene length
  • Solution: divide sequence count for each gene by total sequence count and by length of gene in kilobases
  • Result: RPKM: Rate Per Kilobase per Million Reads

• How can we use RPM and RPKM?

  • RPM: Allows comparison of the same gene between samples
  • RPKM: Allows comparison of same gene between samples, and on top of that, between different genes
  • but underlying hypothesis = debatable (subtle bias