XSEDE 2016 Medicinal Drugs Project submitted by Brian Bartoldson, Eitan Lees, Ian McCann, Amirhessam Tahmassebi and Alex Townsend Reviewed by Wolverine Here are some comments/suggestions on the project, not necessarily in order of relevance or appearance. - Perhaps it could be useful to add, in the introduction, a final sentence where you explain what will be discussed in the following sections. This helps the reader to decide what he is interested in reading and what he can skip. - The conclusions section should be the last one, and discuss what you discovered in the whole paper (including the 'bonus' part). - I would avoid referring to the 'bonus' section as 'bonus'. Perhaps you can use a more scientific and appealing description, such as 'analysis of model enhancements'. Sure, this was a problem that was given to you as an excercise, with a core part and a bonus part. But it is good practice to always write your paper as if it was meant to be published and/or accessed publicly (and perhaps without access to the statement of the problem). - Consider, for instance, the sentence 'Excretion of the drug is the drug excretion rate times the plasma drug level; intestinal medicine/drug level is the prior intestinal drug level minus the drug absorption, plus the drug intake; and plasma level is the prior plasma level minus the drug excretion, plus the drug absorption.' For a scientist/engineer it is often much easier to read an equation than to read a paragraph. Often scientists skim through a paper and try to capture the message from equations and figures. It is much faster. Instead of this paragraph, one could write something like E = r*p I = I_0 - A + u ... where E is the excretion of the drug, r is the excretion rate, p the plasma level, I the intestinal drug legle, I_0 the prior intestinal level, A the drug absorption and u the drug intake. The equation sticks in the reader mind even when he/she is reading two pages ahead. - Try not to fragment the flow of the sentences. For instance, the sentences By manually adjusting the model parameters for number of doses per day and the total daily dosage level, we discovered two distinct relationships. The dosage level controls the steady-state blood plasma drug concentration. The number of doses controls the frequency and amplitude of oscillations in the blood plasma drug concentration. The period kind of 'closes' a self-contained unit, a single thought. However, your thought is not complete after the first period (one is waiting to hear what the two distinct relationships are. You can use the colon: By manually adjusting the model parameters for number of doses per day and the total daily dosage level, we discovered two distinct relationships: the dosage level controls the steady-state blood plasma drug concentration, while the number of doses controls the frequency and amplitude of oscillations in the blood plasma drug concentration. or you can use some connectors: By manually adjusting the model parameters for number of doses per day and the total daily dosage level, we discovered two distinct relationships. One one hand, the dosage level controls the steady-state blood plasma drug concentration. On the other hand, the number of doses controls the frequency and amplitude of oscillations in the blood plasma drug concentration. Either way, do not want your paper to look like a list of sentences. You have to *create a story*, keep the reader engaged between sentences. - Careful when you say something like "the optimal dose is...". One could easily object "have you tried all possible combinations?". Of course that is not possible. In math one can also prove that a solution is *optimal* by showing that it satisfies some conditions (e.g., a maximum of a differentiable function in 1d occurs where f'=0). If you cannot *prove* that your solution is *optimal*, use a different wording. For instance, from your paper it looks like you tested 4 combination of dosages (figure 1). Then you could say that 'among the scenarios tested, the one with 4 doses of 12000mg appears to be outperform the others'. In general, try to avoid *absolute* statements, unless you can support them with evidence and/or proofs. - When you use variables (which is a good idea, especially if you also use equations, see the first comment), always specify what they are. At the beginning of section 3.1 I see log(n), but n is never defined in your paper. If its definition is "short", add a definition, otherwise, refer to another source (adding a citation) for its precise definition (but still say two words about its meaning). For instance: - blabla log(n), where n is the number of doses. - blabla log(n), where n is a quantity that roughly measures the amount of drug in the plasma (see 'Some cool paper', by I. Man and C. America for more details). - Use readable font sizes in your graphs. And don't be afraid of having 'thick' lines. If your paper ends up in a book, it may be subject to a rescaling, and numbers/lines will get even smaller. Make sure people do not need a new pair of glasses to read... - When discussing sensitivities, try to keep in mind that some parameters *cannot* be changed. This does not mean that it is not interesting to do a sensitivity analysis for such parameters, but the conclusion is different. In particular, if the parameter A can be tuned (e.g., the number of doses), then the sensitivity tells us how our actions (the number of doses) change the output (the drug level). On the other hand, if a parameter B cannot be tuned (e.g., the excretion rate), then the sensitivity tells us how different scenarios (different drugs) will behave differently. So consider your very last sentence Lastly, it’s important to note that lowering the excretion rate leads to higher steady state concentrations, which could be more toxic depending on the drug. The word 'lowering' suggests that you can actually change that parameter, while in reality that is fixed (that is how the drug works). Instad, use a wording that suggests how we should take different actions depending on the drug property: Lastly, it is important to note that a lower excretion rate causes a higher steady state concentration, which, depending on the drug, may possibly lie above the toxic threshold. - In your conclusions paragraph you only mentioned 2 findings: the steady state level being correlated to the dosage level and the oscillations being related to the number of doses. However, in section 2.3 there were other interesting findings, that probably deserved some more light (either in place, at section 2.3, or in the conclusions). In particular, you noticed how the half life of the drug impacts the feasibility of a 16 hour window treatment, and you discussed the limits of the model. These are all important aspects. For instance, you could spend a few words on how you think the results would change if we removed some assumptions, such as the 100% absorption. Clearly, you should support your 'guess' with some evidence from the graphs or some consideration on the governing equations.