Update
author |
Steve Losh <steve@stevelosh.com> |
date |
Wed, 06 Sep 2023 11:22:07 -0400 |
parents |
97111cd8535b |
children |
(none) |
# Biology Notes
## 2021-04-26
Homework:
* Watch Ken Burns: the Gene.
* Go through <https://www.coursera.org/learn/introduction-genomics>
* Find practice exams from biology courses.
* Find a paper to go over.
Started Introduction to Genomic Tech class on Coursera.
In one of the videos he talked about the word gene. What does it mean?
* A heritable unit of the genome?
* A sequence of the genome that codes for a protein?
* What about introns? Are they part of the gene?
* What about the upstream region with enhancers/repressors/promoter? Aren't
those just as important as the content itself, since they manage whether the
gene gets transcribed at all?
* What about parts that code for non-protein-making RNA (TODO what is this
called again?)?
## 2021-04-27
More Coursera.
* Any given gene might have a bunch of exons that get spliced in many different
ways. Is there a way to tell just by the sequence which stretches are exons?
I.e. is there something like the start/stop codons for exons/introns?
## 2021-04-28
More Coursera. Learned about some sequencing variants I hadn't known before:
* CHiP-Seq is used to figure out where repressor/enhancer proteins bind to the DNA:
* Start by "cross-linking" the protein to the DNA (basically "freezing" the proteins to the DNA so they can't detach).
* Fragment the DNA.
* Use antibodies (how?) to pull/separate out the fragments that have a protein fused to them.
* Sequence just those fragments to find all the stretches of DNA that have proteins (presumably repressors/enhancers) bound to them.
* Bisulfite sequencing is used to figure out which C's are methylated:
* Start with two identical DNA samples (unsure why we need this).
* In one sample, do some chemical magic that converts all *non*-methylated C's into U's.
* Sequence both samples.
* Any remaining C's in the modified sample must have been methylated.
## 2021-04-29
Coursera CS section. Mostly review for me, but a couple of interesting tidbits:
* I liked how he emphasized how you have to *understand* what the software you
use is doing, and not just treat it as a black box.
* The example of RNA editing and misalignments was interesting.
* Need to look into "NCBI Genome Workbench" program. How does it compare to IGV?
* The example RNAseq pipeline he used was bowtie → tophat → cufflinks → cuffdiff
which was what we used in my class at RIT.
## 2021-05-01
Coursera statistics section:
* I like the term "ridiculogram". I see these a lot.
* Batch effects are very common and very problematic.
## 2021-05-03
Lesson.
Gene definition: a region of the genome with some function. Alternatives might
be called "genomic elements".
Introns/exons *might* use epigenetic information to determine splice sites.
Search for "regulation of RNA splicing" to find more.
Ligand: a molecule that can interact.
Homework:
* Might want to do DNA Algorithms class on Coursera.
* Do at least one of <https://ocw.mit.edu/courses/biology/7-012-introduction-to-biology-fall-2004/exams/>
* Look into supplementary material to find out if we were used in the journal club paper.
## 2021-05-09
Finally getting to the practice tests and the Coursera course. This week has
been crazy.
Started with practice test 2. Some parts were easy, some others I vaguely
recall but am going to need to use the book to refresh myself on.
## 2021-05-16
Trying the scavenger hunt Emily gave me.
There seem to be a LOT of results. Maybe I should try limiting them by date?
Tried to figure that out, looks like there's a way to do it with an "Entrez
Query". <https://www.ncbi.nlm.nih.gov/books/NBK3837/> is the handbook that
describes those queries. I think it should be something like:
2000/1/1:2020/5/5[Publication Date]
Not sure if there's a standard place to find the location, or if it's always
just randomly mixed into the text and you have to figure it out.
## 2021-05-19
Lesson.
Chatted about the COVID scavenger.
Question from last time: how does DNA replication terminate? It goes until the
telomeres every time — it never partially replicates.
## 2021-05-23
Caught up on the coding side of the Coursera course. Redid the gnuplot
interface in my Lisp utilities to match what the gnuplot book recommends, and
it's working pretty well.
## 2021-05-24
Lesson.
CD34+ cells: CD34 is a protein present only in young stem cells, used very often
as a marker to identify them.
Flow cytometry:
1. Pick 1 or more markers (proteins on the surface of a cell).
2. Buy antibodies that will bind to those markers, which have a fluorescent tag
tag attached to them.
3. Draw blood.
4. Lyse the red blood cells with a special reagent.
5. Mix in the antibodies. Antibodies bind to the cells of interest.
6. Use a microfludics thing to pipe cells one at a time through a channel.
7. Use lasers to excite the fluorescent tags.
8. Measure cell counts (and abundance of the marker on the individual cells!).
## 2021-06-01
Flow cytometry presentation. FCS is a super interesting tool that I didn't
really know about before. Interesting tidbits:
* About grill/refrigerator sized, $100k to $200k+ (depending on number of
lasers, more = better).
* You can not only count cells, but also *sort* them on the fly.
* It analyzes basic non-wavelength stuff like the size of the cell (by shadow
size) and density (by how much light is scattered at various depths in the
cell).
* It also analyzes wavelength-specific stuff to figure out which
receptors/antibodies got dyed.