I'm in a similar boat. I have a bachelors in computer science, 13 years in industry, and I'm fascinated by biology.
The other comments are very good, so I'll add things I didn't see mentioned:
* In my opinion, you have to get beyond the wow factor and figure out what you want to work on and why, rather than the how (e.g. CRISPR). I spent half a year researching this and my result was an interest in heart disease [1].
* Credentialism in biology is high, as mentioned. I'm currently in an M.S. in General Biology program. I was able to get my work to shift me to part time (20 hours/week) and I found a non-thesis M.S. program which is geared towards part time professionals (a rare program format) which is perfect for me. However, another comment mentioned biochemistry rather than general biology and there is some merit to that point, although I've appreciated the broader perspective of a general program.
* If you take the Biology credential path of M.S. or PhD, unless you go the Bioinformatics route, you'll likely need to take the GRE Biology specialty test, which was quite difficult for me with almost no biology background. It took me about 6 months to teach myself [2]. I liked the textbook "Campbell Biology" by Urry et al. as well as others, although I find the lack of citations in textbooks really annoying for the way I learn (following rabbit holes).
* Personally, I'm trying to avoid the bioinformatics route because I don't want to just be a tool of some other scientist, but I want to be a scientist myself. This is a hard path and there is some merit to others' comments about joining a biotech startup as a programmer and then transitioning to deeper biology.
* In my networking, I see a lot of people going the other way: from biology to computer/data science. Anecdotally, this seems to be largely due to points others raised: lower salary, a glut of PhDs, slow growth and conservatism of the field due to regulation, etc.
I don't know where I'll end up exactly, but I've thoroughly enjoyed the ride and I think you should scratch your itch if you can. Feel free to email me (email in profile) with questions.
I've become bored with software. I just started a masters degree in Biology and have a vague and naive plan of becoming a cardiologist and running a lab with the primary goal of reducing human suffering, rather than money (been there, done that).
I see, so basically instead of intuiting a simple threshold (e.g. >X% change), they apply an SVM which is able to discover more accurate thresholds (and error ranges). Do you have any suggested resources on learning more about SVM?
I guess my question comes from the observation that these advanced statistical techniques such as machine learning haven't been around for long and yet medicine has often created decision boundaries, presumably just looking at the data and making a reasonable cutoff. Is all the extra effort in a case like this worth the time investment?
Why did they need machine learning? It seems from Figures 2B-G that there's a clear cut off.
"Moreover, to create a classifier for ME/CFS patients capable of identifying new patients, required for a robust diagnostic tool, we developed a trained kernel Support Vector Machine (SVM), a supervised machine-learning algorithm, using our experimental data. To classify new patients based on whether they fall to the right of the decision boundary, we initially selected the two features with the largest significance: change from the baseline to the plateau and change from the minimum to the plateau for the in-phase components of the impedance. Using these features, a cubic polynomial kernel SVM was able to classify the two populations, although the two features are highly correlated, as shown in Fig.2H."
I tried to let the figures and tables float but I disliked the strange spacing that resulted. I also tried putting them at the top and it was very hard to follow.
Sorry for the delay; a few unexpected things came up. Thanks again for the thoughtful comments.
> Most important: Nothing goes into the main body unless it makes a tangible, useful & clear contribution.
> Some of the best papers are the shortest ones: I can read, understand and explain it to someone else in 1 hour. Among the worst are the 25 pagers that take me 2 days to realise that they aren't useful to me.
This is an interesting point because I see both sides of it. Brevity can lead to opacity in some cases; however, I agree brevity's a good ideal (in whatever sense brevity is reasonable).
> - viii & xiii @ on page 3, xvi @ page 4 should either be in main body (if important, they don't look it), or deleted.
Agreed, I'll move those to the main body.
> Things like python commands should just get dumped in the appendix. No footnote / reference. A blanket "you can see all the commands used along with descriptions in the appendix". I'm going to look at what is in the appendix anyway. Because I'm a pedantic academic.
Agreed, I'll move the commands to the appendix.
> - Think about using a template like [1].
> - Reduce your font size to 10pt please.
> - If you are going to be very maths heavy, think about moving to a single column style. I, personally, find it makes it easier to read eqns and to follow their logic. Currently I have to jump from around the page and keep getting lost.
I do like the double columns but I'll think about this.
> - For sections 5 & 6: Stop putting something in bold every paragraph. Bold is only to highlight when it's really important. The name of something is not really important. Prefer italics over bold, but even use that sparingly. How difficult was it to focus on reading this paragraph when the words keep changing shape?
I didn't think about that. I'm not sure why I decided to do that, but I agree and I'll remove those.
> Graphs & Tables exist at the top of a page. That's the only place they live if they live in the main body. They don't have to be on the same page as where you refer to them, and you can group them together. Otherwise, appendix that stuff.
> Else you'll to end up with blank space (like end of page 4 & 5) and formatting headaches later on.
Yeah, I had to use `\raggedbottom` for a nice flow. I'll try the more classic way.
> You haven't done your results discussion part yet... When you do, make sure you don't talk about every single table/graph. Only talk about the results that are important. Otherwise it's guff that will bore your audience.
Makes sense.
> You have so many sections that I need a table of contents to work out where I am. For a 10-15 pager, that's silly. Learn to love subsections. Especially those early parts.
I hadn't considered sub-sections. I'll try those out.
> I have no idea what previous work this relates to. Is there previous work in the field? If so, talk about it. Talk about how you're improving it. Talk about what the context of this paper is.
> Don't know what the context is? Then you better find out... People will ask!
As far as I could tell, not much, or very orthogonal, but I've seen this feedback multiple times, so I'll do my best.
> This seems like a technical paper, not an English lit assignment. If you directly quote anything, let alone a whole paragraph, it better blow my mind. I am afraid page 3, column 2 does not. Remove it. Just reference anything like that. If people want/need to know, they will read it too.
That makes sense and I'll remove that huge paragraph. I think it's worthy of highlighting, but best for the appendix. I do wish academic writers would more often quote what they found most relevant from critical papers instead of just using obscure citations.
I'll be away for a few days, but I'll integrate this feedback soon. Thank you very much, again. I'll post an updated draft here when finished, or feel free to email me (see my bio) if you'd rather get a push then pull (and I totally understand if I never hear from you again). If you ever need help with anything, please let me know.
The conversation is probably mostly a part of the field of Positive Psychology and Meaning in Work literature, a summary of which I cite (Martela and Steger); however, when researching the field (I probably read about 50 academic articles), I didn't see any discussion about value theory (in the realm of philosophy). I have a connection to one of the researchers and maybe I can reach out to start the conversation.
That's a very helpful summary, thanks. From another comment, it seems arXiv requires academic affiliation or endorsement: https://arxiv.org/help/endorsement
I'm not really sure what field I'm targeting, so I was hoping people would note their favorite open access journals and I could investigate (and I also thought a more generic list would be generally interesting to all the HN folk).
For those looking for a straightforward introduction to forecasting, see the free e-book, Forecasting: Principles and Practice by Hyndman and Athanasopoulos: https://otexts.com/fpp2/
The other comments are very good, so I'll add things I didn't see mentioned:
* In my opinion, you have to get beyond the wow factor and figure out what you want to work on and why, rather than the how (e.g. CRISPR). I spent half a year researching this and my result was an interest in heart disease [1].
* Credentialism in biology is high, as mentioned. I'm currently in an M.S. in General Biology program. I was able to get my work to shift me to part time (20 hours/week) and I found a non-thesis M.S. program which is geared towards part time professionals (a rare program format) which is perfect for me. However, another comment mentioned biochemistry rather than general biology and there is some merit to that point, although I've appreciated the broader perspective of a general program.
* If you take the Biology credential path of M.S. or PhD, unless you go the Bioinformatics route, you'll likely need to take the GRE Biology specialty test, which was quite difficult for me with almost no biology background. It took me about 6 months to teach myself [2]. I liked the textbook "Campbell Biology" by Urry et al. as well as others, although I find the lack of citations in textbooks really annoying for the way I learn (following rabbit holes).
* Personally, I'm trying to avoid the bioinformatics route because I don't want to just be a tool of some other scientist, but I want to be a scientist myself. This is a hard path and there is some merit to others' comments about joining a biotech startup as a programmer and then transitioning to deeper biology.
* In my networking, I see a lot of people going the other way: from biology to computer/data science. Anecdotally, this seems to be largely due to points others raised: lower salary, a glut of PhDs, slow growth and conservatism of the field due to regulation, etc.
I don't know where I'll end up exactly, but I've thoroughly enjoyed the ride and I think you should scratch your itch if you can. Feel free to email me (email in profile) with questions.
[1] https://github.com/freeradical13/ValueBasedPrioritization/ra...
[2] https://freeradical13.github.io/