Flora just refers to the plants of an area (e.g., Flora of Singapore).

It's difficult to say whether a frog is more closely related to a gecko or a fish, but it's likely that it's more related to a gecko as it shares a more recent common ancestor.

The PIR-A receptor is shown by pathogen-derived antigens.

This is because red blood cells have a surface protein called CD47. This is the so-called "don't eat me signal".

In the case of malaria, the plasmodium infects the red blood cell. In this case, the plasmodium's antigens are displayed on the surface of the blood cell - if we have antibodies against the antigens, then the antibodies' Fc receptors can attach to the antigens.

This organ also has Langerhans cells - this cell can migrate into the dermis (i.e., a deeper layer of the skin) and differentiate into immune cells.

TL;DR... it's to help see if a decision is viable or not.

The founder of Moderna suggested two things:

1. Nucleoside modifications could prevent the RNA from being degraded in the body.
2. The structure of the RNA could be altered so that it cannot be degraded by the human body's enzymes.

The main point behind this graphic is that a bio-business stems from a technological advancement of sorts that could potentially solve an existing problem.

This "advancement" could then be used in various ways - for instance, the environment, food security, human health, and even the industry or a combination of the said fields.

The effective population size is the amount of individuals in a species needed to experience same amount of genetic drift and or inbreeding as the actual population size.

One thing to note (from ChatGPT) is this:

The more positive the F3 score, the likelier it is that introgression happened between the species being tested (i.e., species "W" in this case) and the first species (i.e., species "A" in this case) and vice versa.

Likewise, a positive "D" statistic would indicate gene flow between "W" and "A" and vice versa.

There are four main sources of evidence that weren't super obvious from prof. Jarkko's lecturing slides, but here they are:

1. ab initio predictions: this is super useful in cases whereby there is limited information available about a gene. We can use a HMM to predict its introns and its exons.

2. Homology-based estimates: here, we look at the gene models and the assembly of similar organisms and apply the guilt-by-association principle to deduce their functions.

3. ESTs and Raw RNA Sequencing Data: this just refers to data that comes straight out of the wet lab - for instance, a ChIP-seq experiment to provide real time evidence of protein-DNA interaction.

4. Comparative genomics: here, we compare an assembly against reference species to find conserved genes, regulatory elements, and also synteny (i.e., syntenic blocks).

This is a gene predictor model that takes in multiple things:

1. ab initio predictions: Augustus relies on a machine learning model called a "hidden Markov model" for this. This model is then used to predict gene introns and exons based on their nucleotide sequences.

2. ESTs and RNA Sequencing Data: Augustus relies on RNA sequencing data from a reference source too to improve gene prediction accuracy (as the data provides evidence of transcribed regions).

3. Training data on species-specific or closely-related species: going off the point above, Augustus can also take in training data that is specific to, if not closely related to the species of interest.

Augustus goes through proceeds through the gene features like so:

1. Input data is given and depending on the training data provided, Augustus can be trained to only predict genes from that specific species.
2. Augustus then uses the above gene features (i.e., the ab initio predictions and whatnot) to predict the locations and the genes in the assembly.
3. Augustus then applies scoring and filtering to build probable gene structures given the models and the evidence.

The above results in an output that contains gene locations, exons, introns, and sometimes, confidence scores for each prediction.

An important thing to note in this graphic is that one molecule of 3-phosphoglycerate and 2-phosphoglycerate each are made.

This is from a dicot.

In other words, one also has to rely on phylogeny for this bit.

This should say exactly once.

With this, you want to aim for a coverage of 20x to 30x. A coverage of 16x is needed for re-sequencing.

A "ruderal" is just a plant that thrives in poor growing conditions (e.g., infertile soil) - so, something like a pioneer species.

This should say "endoplasmic reticulum".

This is the family of viruses that contains viruses like poliovirus.

This is a pretty high concentration as the usual protein concentration is in the picomolars.

This means that the \(\beta\)-amyloid can cause other proteins to turn into \(\beta\)-amyloids.

One thing to note is that even though a protein is folded, it is in dynamic equilibrium with the unfolded states.

What this basically means is that the protein can still become unfolded.

Just FYI - an ortholog is a conserved gene across multiple species.

Or in other words, HBS3 interferes with the MAP module that Ras protein depend on to induce downstream signalling effects.

Basically what this is saying is that if we have KRas (i.e., a kind of Ras protein) and that Ras protein has a G12C mutation, then we can target a small pocket using small molecules to irreversibly keep it inactivated.

Cdc42 belongs to a class of proteins called Rho. However, it's able to bind to GTP very well (i.e., with high affinity).

One thing to note is that a ternary complex forms during the attachment of SOS. It's this complex that allows the H loop to exchange GDP with GTP later on.

There are actually four enzymes that are involved in the post-translational modification of the Ras protein.

1. RCE1 removes that "AAX" group.
2. ICMT1 changes that CO_2_ to OCH_3_

Another thing to note is after Ftase does its thing, the GDIs sequester Ras to the cell membrane where RCE1, ICMT1, and PTase can do their stuffs to anchor the Ras protein to the membrane.

One thing to note about Ras is that it's also not a super fast GTPase.

The thing about viruses like this and the polyoma virus is that they're examples of double stranded DNA viruses. They use early gene expression strategies to induce the infected cell to enter the S phase of the cell cycle.

However, this induction also interferes with the usual mechanics of the Rb protein.

One thing to note about this gene is that it's also a co-factor even though its function is to inhibit transcription in the G1 / S phase of the cell cycle.

The thing about lactate is that an enzyme called lactate dehydrogenase takes it and turns it into another product called pyruvate. In doing so, NADPH^+^ is made.

Basically... both copies of the Rb gene (i.e., on 13q14) need to be inactivated for Rb to be inactivated.

One thing to note is that tumor suppressor genes can only ever be inactivated by recessive mutations.

This should just say "reciprocal".

Note that amino acid - Glutamine 61. This is super important for this reaction!

GAP is super important as an enzyme - Ras is a super lousy enzyme for hydrolyzing GTP (as this reaction has a high activation energy)! GAP solves this by speeding up the slow parts of the reaction by doing stuff like cleavage, phosphorous release, and isomerization.

This is another way that the Sev gene can be unexpressed (i.e., inactivated).

KDM is responsible for demythylating histones so that they can be transcriptable.

This is super bad because...

1. This can overexpress of cancer genes.
2. This can suppress tumor suppressor genes.

Although R-2-HG can be turned back into \(\alpha\)-ketoglutarate, this isn't a super favorable reaction.

The "WT" is the normal type of the protein; everything else are mutants. The bar graph shows that the other mutants of IDH1 have very different (and poor) abilities to reduce isocitrate to \(\alpha\)-ketoglutarate.

The "kM" in the table refers to how well a substrate binds to an enzyme (i.e., affinity). Lower kMs mean that the substrate binds tightly to the enyme and vice versa.z

The "Vmax" refers to the reaction velocity - how fast the reaction is happening.

Red dot in the picture is an added safety feature. Furin is a protease that's found in all cells.

Two kinds of cell types for this:

1. HEK cells

   Very easily transfectable - super efficiently transfected with DNA.

2. MDCK

   Growing fluid.

Ideally, we'd mix both those cells and transfect them with plasmid for amplification. Then, we purify the cells and do some sequencing.

Once all of this is done, then use those two cell lines again to amplify the plasmid.

Four different plasmids are needed to reverse engineer negative stranded RNA viruses - one for the viral RNA, one for "L", one for "P", and one for "N".

+ve RNA viruses rely on DNA because...

1. It's easier to mutate
2. It's a lot more stable than RNA

In other words, if you were to just get that RNA from the virus and put that into a cell, an infection would happen.

Note that most cancer cells let the pyruvate turn into lactate into the cytosol.

This is because Rb binds to a protein called APC/C - this is a ligase.

The main point of this table is that the Rb gene is involved in many other cancers too (i.e., not merely eye cancer).

Matrix-metallo proteases are a good target for anti-cancer drugs as this is a crucial protein for tumor growth.

The studies suggest that one's risk for cancer goes higher the older they are.

In other words, how likely you are to get cancer is NOT independent of age.

In other words, the tumor is made of cells that either have their X chromosomes inactivated or activated.

If one grows cells with Philadelphia chromosomes and perform sequencing with them, they'd find the same breakpoints as they would with a normal cell.

This means that the cancer can only happen if and only if the translocation happens (which is support for the clonal origin of cancer).

There are also two more scenarios to consider: increased cell division and decreased cell death (i.e., apoptosis).

Too many cells leads to higher probabilities of tumors occurring (i.e., cancer).

This is also called the Ewald sum.

Basically put, it's a faster method to calculate electrostatic forces. The "trick" that Dr. Mu is talking about requires splitting that \(\displaystyle \frac{1}{r}\) into two parts.

In other terms, a "score" is calculated based on how well a ligand binds to something.

The above is part of a virtual screening setting. In this scenario, a scoring function evaluates predicts the ligand's binding affinity and based on the latter, assigns a numerical score based on various parameters (e.g., van der Waals forces, electrostatic interactions, etc).

This score is then evaluated using various metrics - for instance, an AUC value from a ROC curve.

In other terms, the temperature in an MD simulation is dependent on the kinetic energy of the atoms in the simulation.

There are many PBCs that one can use, but a box one is the most obvious (and also the most intuitive). However, a box may lead to inefficient computing times - depending on the shape of the protein being studied, the box may be virtually empty.

A good rule of thumb is to keep the box at least 10 Angstroms away from the solute. In a new system, there should be a margin of at least 13 Angstroms between the solvent and the box.

To avoid short range artifacts, the periodic box vector should be at least twice as big as the cut-off radius.

AMBER is short for Assisted Model Building with Energy Refinement.

This is by far the most common. There are other combination rules though...

Most terms in a force field will be governed by an equation that is similar to this.

This is a super equation to know!

The Morse potential is a formula that estimates the vibrational (i.e., potential) energy in the bonds between two atoms.

Or in the context of a Molecular Dyanmics simulation (i.e., MD simulation), a force field is just a means of propagating the system's atoms' positions in each time step of the simulation.

This "energy" refers to the potential energy!

When dealing with a drug and a ligand, we only consider the drug and its binding pocket.

This reflects the "dynamic state".

This value is not time-dependent. \(r_t\) can be an atom or a group of atoms.

This is also most suitable for describing large proteins. When a protein is globular and small, the value of \(R_g\) is small.

Oncology / hematology can be pricey due to the difficulty in manufacturing new drugs. Also, demand for new drugs in this sector is inelastic.

"Chronic" implies that consumers are more sensitive to price changes.

"Acute" implies that the person in question is more willing to pay more (i.e., price-insensitive).

"High" means that something is common and there is a lower price.

"Low" implies that if a disease is rare, then the price will be high (as there are no available substitutes).

Patient age is as such possibly due to reasons such as lifestyle factors and different pricing rates for different age groups.

Mode of administration is also something interesting to look at - possibly due to one's perception as an example.

Prices are a push-and-pull situation between companies and markets.

The company will set a minimum lower limit that the antibody can sell for (to pay back investors, etc). The market will set an upper range that consumers are willing to pay to use the drug.

(A) Focuses on the USA and Europe. More drugs appear to be approved over time. The amount of biosimilars are also increasing as the pipeline is getting better in addition to general knowledge (i.e., patents expiring).

Biosimilars are compounds whose active components are the same, but the formulation may be similar (i.e., slightly different).

(C) Shows the percentage of production systems. CHO cells occupy a large proportion of production systems - there is a market trend towards this. The majority of antibodies are human antibodies; the second largest one is humanized.

(E) Suggests that certain companies are producing most of the therapeutic mABs.

(B) Implies that cancer (i.e., oncology) and dermatology (i.e., skin conditions such as psoriasis and dermatitis) are the most common disease targets for antibodies.

(D) Shows that more and more antibodies are becoming humanized.

(F) Suggests that the USA and Europe are manufacturing hotspots. Singapore is also a hotspot for manufacturing mABs - Japan and China are two other major manufacturing hubs in Asia.

It's the same for every business (in a general sense). Sales are sales.

People (of any kind) can also choose to report any side effects of a drug. Adverse effects can come from lab experiments too.

The agency may - using the above information - choose to alter a drug's safety notifications (i.e., "this drug is not suitable for so and so", etc).

This isn't an actual sampling method, but to speed things up.

This should just be "sampling"

Open Babel is also another great software.

Scoring is super critical here.

This is also problematic as this implies that the protein in question changes its conformation.

We don't know how the protein will change!

There is also something called phenotype drug discovery - we're only interested in the ligand itself.

However, BS3008 only focuses on structure-based drug discovery.

"Angels" are people who want to invest money into your antibody.

Although it costs about 50 million USD to develop a drug, the real number is this high because of failures during experiments.

"Efficacy" refers to ensuring that the antibody works.

The longer the study in Phase III, the more expensive the phase is.

Each rectangle represents the capture value of a drug.

A bad drug can still net a good amount of value if it comes first because it's one of the only drugs in the market and doctors are prescribing it.

The "International PCT filing" period is a period that more information can be added. This period is also when somebody can abandon a patent. The first 12 months in the figure is for protection purposes.

Publication makes a patent known to others (should they choose to look for it). At the "International Phase", the patent is also discussed to propose refinements and content changes. The patent can also be rejected here.

The cost of the "National phases" is high because each country in the figure will undergo the same country (which country is chosen depends on the market that you want to sell to). The processed patent claims from each country will go back to the lawyer (who will review the claim as well).

Combining multiple mutations can also work here.

This is the most important one to note.

Others cannot use this discovery, but this doesn't mean that the patent holder can use their discovery either.

This is when the patent is filed.

The expiry date can be extended by various means, but this is more of a general scheme.

Generally, academic institutions will give broad licenses (i.e., permission to do something), but never relinquish the entire patent.

In NTU, the ownership is about 50% (but to NTU also controls the final product), but this varies from institution to institution.

These classes are used for small molecules.

With this class, one can also calculate frequencies and vibrations.

These are all boundary interactions. Most force field systems treat these as partially non-bonded.

All of these parameters (i.e., bond stretching, valence angle bending, etc.) are known as hard parameters.

The torsional angle is the toughest to calculate. The degree of freedom of a peptide chain comes from this same torsion term.

Affinity and epitopes are necessary for high binding.

But this is a lot of work, so it's not as common (though it is still common).

The most common is an indirect ELISA method. ELISA is the most common.

Sandwich ELISAs have greater specificities, but two epitopes are needed (which may be cumbersome).

For indirect ELISAs, we also need the antibody's target and also the detector binding.

This could be a salt or a pH gradient. This is used in industry after basic affinity chromatography.

Protein A, protein G, and ion exchange are good purification methods.

The typical components for cloning are:

1. Promoter
2. Ribosome binding sites (where the ribosome binds)
3. Restriction site (i.e., RS1) - an enzyme that will allow you to cut and paste a gene into another one.
4. A sequence of your design, but prefaced with a signal peptide (e.g., ompa).
5. A VH Domain and a CH1 Domain (linked by glycine-serine linkers)
6. Histidine tags for purification purposes
7. A STOP codon

Antibodies are generally produced in the periplasm - this area is safe for disulfide bridge formation. Bacteria can only produce Fabs or single-chain Fcs. The antibody is secreted into the medium regardless of animal or bacterial systems.

The most common way to do Fab fragment production is to leave the sequences for the VH and VL elements with more DNA in between. This is because it's easier to do gene synthesis.

The pET vector on the right is a common example.

Higher antibody concentrations lead to lower signals.

Note that "HEK" refers to kidney cells. "CHO" is short for Chinese Hamster Ovary. A requirement for choosing a system is stability.

The forward scatter detects the shape of the cell.

The side scatter detects the granularity of the cell (i.e., what's in the cell).

Between the two extremes, you'd want a "Slow on, slow off" behavior.

In the case of a cancer cell, we'd want the antibody to last as long as possible - this wouldn't be possible if the antibody bounded quickly, but also left quickly!

This is the most common purification method.

In the final step, a substrate is added to perform a color change.

Proteins A and G are bacterial proteins that are associated with infections.

They are used in recombinant technology for affinity purposes. Both proteins A and G bind to the constant domain of the antibody in the heavy chains.

Protein A has some sort of advantage for some VHs. Most companies use protein A purifications.

A column-like device is used for this. Buffer is used to wash other compounds away.

This is more common because the end-product is similar to human systems.

A single or a double plasmid is possible. There are two options for the heavy chain:

1. The "Kozak" is also a restriction binding site. It allows translation.
2. The His tag is optional - it can be changed to something else.

For IgGs - there is a hinge region in the middle.

The same genes in a mammalian plasmid are also present.

It's more difficult to use two plasmids because of the origin of replication genes.

If a bacteria has two plasmids with the same origin of replications, the bacteria may discard one of the plasmids and not produce equal quantities of both plasmids' products.

This is doable with plasmids with different origin of replications (provided that they are compatible with one another).

The graphic on the left should be used to model the torsional term energy for ethene.

This is because ethene is a double-bonded molecule - its local minima is at 0 and 180 degrees.

For ethane, the graph on the left should still be used. When the angle is zero, the hydrogens overlap with one another.

The amplitude of both functions reflect the amount of energy needed for the molecule in question to adopt a different conformation. However, in reality, more terms are needed for more accurate representations.

The \(\displaystyle \left(\frac{\sigma}{r}\right)^{12}\) represents repulsive forces between two atoms while the \(\displaystyle \left(\frac{\sigma}{r}\right)^6\) represents attractive forces.

For more information about the Lennard-Jones potential (i.e., an alternative name for this formula), do check out this link.

The equation for \(E_p(r)\) is based off the equation for the potential energy in a spring: \(\displaystyle E = \frac{1}{2}Fx\).

Since \(F = kx\) (i.e., Hooke's law), it follows that \(\displaystyle E = \frac{1}{2}kx^2\) (where \(k\) is the spring constant).

The variable is still \(r\); the parameters are \(k\) and \(r_0\). This formula for \(E_p(r)\) is also known as constraint energy.

The equation for the potential energy in this system has a name - the Lennards-Jones Potential. The equation has the following parts:

1. \(\epsilon\) is the well depth - the distance from the x-axis to the minimum of the potential energy curve. If \(\epsilon\) is small (e.g., like for Neon - a noble gas), this means that the well is shallow and vice versa. \(\epsilon\) represents the change in potential energy.
2. \(\sigma\) is the distance at which the potential energy is zero. Dividing \(\sigma\) by half yields van der Waals radius of the atom in question (i.e., \(r = \frac{\sigma}{2}\)).
3. \(r\) is the distance between both particles (i.e., the x-axis).

\(r\) is still a variable here - the parameters are \(\epsilon\) and \(\sigma\). Note that the value of \(r\) that minimizes \(E_p(r)\) is \(\displaystyle \frac{dE_p(r)}{dr}\).

The particles cannot move freely in this case. This is because any other positions would result in unfavorable energy states.

At longer distances of \(r\), both particles are too far apart to interact.

At shorter distances of \(r\), both particles are brought closer together and both particles also experience an attractive force. This continues until equilibrium distance \(\epsilon\) when a minimum bonding potential has been reached.

At even shorter distances of \(r\), the potential energy rises because the repulsive forces are greater than the attractive forces.

Both particles are free to move in this system.

The variable is \(r\); the parameters (i.e., changeable variables) are \(q_1\) and \(q_2\).

When \(q_1\) and \(q_2\) have similar charges, the potential energy graph would like a decreasing exponential graph due to repulsive forces.

Potential energy in this case is a function of the coordinates of all atoms.

\(r\) is the Euclidean distance between two particles.

This case deals with the polar charges of water.

The "force" of an atom is the sum of the partial differential of each coordinate - in this case...

\(\displaystyle \vec{F} = \frac{\partial E_p}{\partial \vec{x}} + \frac{\partial E_p}{\partial \vec{y}} + \frac{\partial E_p}{\partial \vec{z}}\)

This also implies that a change in any of the atom's position changes its force too.

When \(\vec{F} = 0\), there is no change in the system. \(\vec{F}\) is the result of potential energy!

This is mostly constant across all systems.

In other terms, although we don't attempt to model the wavefunction directly, we still make models to approximate the system.

For instance, in a water molecule, an oxygen atom is represented by a red bead and the two hydrogen atoms as two smaller beads. However, this model ignores electrons, let alone interactions with other water molecules (we can rectify this using the partial charges model).

Potential energy models are used to simplify things - ignoring electrons. Whatever properties that a molecular has is governed by its potential energy.

These determine the properties of a system (i.e., molecule).

My guess is this:

Assuming that the system in question has more than two atoms (i.e., particles) in a 3D space, we can apply the potential energy graph for each pair of particles that we can form.

Because \(k\) is constant, it follows that for any pair of particles we consider, the coordinate \((x, y, z)\) that minimizes the potential energy will be a constant distance away from \((x_0, y_0, z_0)\).

This is characteristic of a lattice.

The \(q\)s in this equation are the charges of the particles.

\(r\) is the distance between the two particles.

\(\epsilon_0\) the dielectric constant of a vacuum (i.e., \(8.85 \times 10^{-12}\) Farad / meter).