Hana Bashir Flashcards
Wild caught mice/rats/animals are good because
- more representative of entire population the experiment will be generalised to
- variability in genetics, environment, microbiome etc which are confounding variables.
ARRIVE
- Animal Research: Reporting of In Vivo Experiments
- randomisation method must be stated
- power calculation
In mouse models it is
- good to randomise sex, age etc
- if all mice in one treatment happen to be younger, this will confound results
Ethics
ethical approval number required
miRNA effector
insert the target sequence into a protein and see if the miRNA cuts it?
A good study would use both
male and female mice, although lots exclude females as results can be affected by stage in menstrual cycle - this is problematic
Inappropriate techniques to use on small animal models
in-vivo imaging techniques i.e fMRI or PET due to small brain size
Immunohistochemistry
positive control (add dye), negative control (add water)
Yellow/Green fluorescent proteins
non toxic
CRISPR
- can produce off target cuts
- any observed results that seem due to the treatment might not be
retrotransposons often have
multiple copies across the genome
RT-qPCR
quantify gene expression levels
response variables
could be influenced by other factors eg environment
calcium
spikes per minute, visualise the data
Two independent mutants in parallel
- must have the same phenotype
- if you mutate A the first time and get a different phenotype to mutating A the second time then issue must be in experimental design and observed results may not be valid
Generation of 2 independent lines for each translational reporter is appropriate as it
controls for insertion affects (if both lines show the same fluorescence pattern it suggests the fluorescence isn’t just an insertional artefact)
Use of a constitutive promoter is appropriate when
- you want to confirm transformation (if a cell fluoresces you know it’s taken up the transgene)
- general cell tracking (if you want to label all cells e.g tag all neurons with GFP)
C elegans, Arabidopsis, drosophila all have a
short generation time
well annotated genome
- quicker/easier to identify genes of interest and their functions
- know where genes are located so more precise CRISPR, RNAi
- comparative studies
maize genome
- too large to be fully sequenced
- use rice instead
Zebrafish
suitable for studying human disease as have a highly conserved genome to us
Yeast
suitable for studying cell cycle regulation as conserved core mechanisms
takes about a month for
- lab mice to acclimatise
- longer for wild rats (otherwise stress may impact experiment and change the rats’ physiology)
Random allocation into groups is appropriate method but
- randomisation method must be mentioned
- was it computer generated, people picking an envelope
organisms have been divided into groups according to as many variables as possible to control for these, eg, age, sex, etc.
THEN once you’ve made groups you randomly assign which treatment each rat in the group gets (to reduce bias)
Method of single housing the rats is not applicable to real world scenarios
rats pass on information culturally, so if poison makes them sick they may indicate this to conspecifics but this is complexity is missed by single housing
No choice feeding of poison
good because otherwise rats may not eat it but bad because not applicable to real life where they could choose to avoid it.
Not appropriate to fit a super complicated model to a small sample size (overfitting)
- unstable estimates
- unreliable predictions
- model learns noise instead of true patterns
