Is mathematical aptitude the limit of how well I can do in an interdisciplinary life science field as a computer scientist?

Question

_{I apologize in advance if this question is too broad, or too subjective - I couldn't think of a better place to ask it.}

Summary: does mathematical aptitude impose an upper boundary on how well one can do in research areas where life sciences intersect with engineering and computer science? I.e., is an ideal researcher in such fields either an experimentalist or an applied mathematician/physicist?

---

I'm a computer science graduate thinking about pursuing a PhD in an interdisciplinary life science field; specifically systems biology, synthetic biology, bioinformatics, genomics or cognitive (neuro)science. Characteristic of all these disciplines is that they seem to - ideally - require one of the following two profiles:

• someone with great experimental/domain skills (e.g. molecular biology, neuroscience methods etc.),
• someone with great quantitative skills (math/physics).

By virtue of my undergrad background, I'm likely closer to the second profile, even though I have some domain knowledge. However, here's where my self-doubts begin. People of this profile seem to be - in the long term at least - expected to produce new theoretical knowledge primarily by employing advanced math. Top research papers seem to be full of it. I'm concerned about this for the following reasons:

• my mathematical skills and aptitude in the context of this profile are average, or a bit above average at best. Sure, I can handle ordinary differential equations and numerical integration, but once it gets to the postdoc level and above, I'm competing with people who are elite talents and have backgrounds in math or physics from top schools. I don't believe I posses anywhere near the math talent that they have, and I simply wouldn't be able to do the job as good as them - and I don't want to be producing subpar research. Furthermore, I spent a significant amount of time studying things that don't seem applicable to this type of research, like CPU architecture or OS internals. Because I'm targeting programs in Europe - where a PhD is normally 3 years - I can't in this short time develop my math skills enough to close the gap, especially considering my less-than-elite talent (funny as it may sound, I believe I have a much greater talent for the humanities/philosophy, but the amount of positions there is close to zero).
• by far my strongest actual skill is coding/programming, and I could probably get some kind of position based on that. I believe I could do well enough to produce some tools and maybe even get a (likely mediocre) PhD eventually, but after this step - when an independent research path is expected - it seems to me I would be at a disadvantage and heading for the industry, because implementation skills are simply not enough, i.e. they seem to have merely a supporting role in academia (since they alone cannot generate new knowledge, the main objective of research). Permanent faculty positions seem to go out at much higher rates to math-inclined individuals, who are at an advantage when it comes to formulating a unique theoretical research statement. At least, this is my observation.

Should I choose something else if I'm not a top math talent and thus can't evolve into any of the two profiles mentioned?

Or is such categorization of researcher profiles in fact a false dilemma and there are more options, i.e. not everything required to perform well in these fields can be reduced to mathematical aptitude?

For what it's worth, I did undergraduate research in computational biology that lead to published papers, and sure I could handle math at that level, but it just seems to me that to make it as a top full-time researcher, I would either have to become an applied mathematician or an experimentalist.

Answer 1

Speaking as a computer scientist who works with life sciences: there is a lot more use for computer science than just modeling.

Consider:

• Who makes and maintains the instruments?
• Who handles LIMS development and deployment?
• Who develops automation systems and precision protocols?
• Who develops the control mechanisms that actually implement biomedical intervention concepts?
• Who develops taxonomies, interchange standards, and data representations?

This just scratches the surface of computer science opportunities in the biomedical Unfortunately, the view that you hold of "you're either an experimentalist or a modeler" is far too prevalent in biomedical culture, and computer scientists often get little respect for their contributions. That culture is starting to shift however, and especially in younger and more engineering-centric subfields like synthetic biology... and especially in some of the awesome things the younger practitioners are doing.

Answer 2

From your question and comments, it seems you view “success” as being involved in original research, in a long term position where you can focus passionately on your research.

You are less concerned about success in completing the PhD program, and more concerned about “10 years down the road” from that, about the research, about being involved in compelling original research, on an ongoing basis, being in the thick of exciting things that are being worked on.

You are concerned that a wrong choice NOW would lead to a poor outcome, where you would wind up doing less-interesting research, a career, a life, of frustration, preventable by a better choice up front. (Can’t say I blame you)

You have observed two types of profiles that succeed in this field: 1) Someone with great experimental/domain skills (e.g. molecular biology), 2) Someone with great quantitative skills. Of these, it seems you pretty much write off 1), apparently due to lack of background, and focus on 2), with the observation that you have limits in this area. So something closer to the great quantitative skills is how you feel you’d make your mark.

You feel your current advanced mathematics skills may limit you. I also infer that you do not wish to fully give your mind over to the advanced mathematics area, as it does not excite you as much as other areas.

So you are looking for an answer that says, "Yes, you can make this work, and here's how, and here are examples of people who have done this."

You mentioned your concern about matching up to the mathematicians, competing with them. “I'm competing with people who are elite talents and have backgrounds in math or physics from top schools. I don't believe I possess anywhere near the math talent that they have, and I simply wouldn't be able to do the job as good as them - and I don't want to be producing subpar research.”

I challenge your assumption that you are inherently and always competing with these people. Why don’t you focus on collaborating with them instead, certainly at least at first? Why not see if you can provide all the key skills around this area so they can collaborate and plug in with their terrific skills, their 10%, in service of your great experiment design? In the process, you will see how they think, your mirror neurons will get busy helping you learn rapidly, and you can decide if you want to learn to do what they do, or if you are just as happy to let them do that while you address parts of the problem that feel more creative and interesting to you.

I suggest that a good next action step is to start by looking again at some of those many papers you mentioned that contain large quantities of advanced mathematics, and then check the author list, and use your own research skills to ascertain which co-authors were responsible for which parts of the paper.

If possible, contact some of them and ask--how did they divide up the mathematics portions of the paper? Do some of the collaborators even possibly have (gasp) a similar level of mathematical skill to what you currently possess, but great skills and creativity in other areas that were critical to success? You might also make some terrific contacts and have some amazing conversations in this process.

I think the answer is to go back to the papers that had the advanced mathematics that looked intimidating, and do some data-gathering fact-checking on which authors really did what, and use them to get a more accurate, and likely more diverse, profile of the types of people who can be successful in this field, and with any luck at all, potentially find a few models you can identify with, and inspire you to take the plunge. (Who knows, maybe in the process you will discover other papers that are equally interesting to you, with less reliance on the advanced mathematics.)

P.S. If, in the process, you don't find ANY models that inspire you, give you confidence, then DON'T DO IT!! Run Forrest, Run!!

Answer 3

This question doesn't seem to be active anymore, but I just came across it and there's one thing that nobody has talked about: big data.

Data is everywhere in the "wet science" nowadays, and it really is changing the way people do research. I would say data analysis skills are at least as important as pure math skills, and these skills are a mixture of intuition, good programming skills, understanding of data management, statistics, machine learning, and perhaps more.

Good programming skills are also not very common in the wet sciences. My friend who's a glaciologist struggled with that in his PhD, doing experiments over and over because he didn't trust the code he wrote, and when they hired a programmer, the programmer didn't understand the domain, so he was pretty useless.

The opportunities of applying good data analysis skills are huge, even in the humanities. There's a major international challenge called "digging into data" shared by the US, Canada, the UK and the Netherlands that really recognized that (http://diggingintodata.org/).

Answer 4

It seems to me you are making excuses and "failing yourself", i.e., declaring your own failure in advance, for reasons that may or may not even wind up being relevant. I read in your question that you do in fact actually WANT to go for this, but you are afraid of failing in the end. So, the real question becomes, how do you define your own "success" or "failure" in this endeavor? If you wind up getting a PhD at the end of the day, will that in itself constitute success? If you pursue this path, but do not obtain a PhD, or do not obtain a permanent faculty position, will one of those unequivocally constitute "failure" in your own eyes? Is it still a failure if you wind up obtaining well-paid employment doing work you are passionate about, with brilliant people you love working with?

Are you trying so hard to have everything "perfect" according to the limits of what you currently know, that you are keeping yourself from even pursuing something you appear to want?

It seems there are a range of possible outcomes here.

1) You don't even get selected to start in the PhD program, and the whole issue is moot.

2) You enter the PhD program, and in the course of your studies, discover some other, related area you are passionate about and particularly good at, and change your studies and degree program to focus on that previously unknown area.

3) You enter the PhD program, do not enjoy it, do not complete it, and discover other areas of life you wish to devote your energies to.

4) You enter the PhD program, and complete it, but not at the full level of distinction you wish, and at the end, wind up having less ideal choices than you would like to have, though one or two of the choices are reasonably appealing.

5) You enter the PhD program, and complete it, in the process gaining all the advanced math skills you truly need, bringing your own unique skills to bear, and completing the degree with distinction, and have numerous appealing options to choose from at the end,

This is of course an artificial, limited list of the range of possible outcomes. Which of these would be success? Which would be failure? It seems to me entirely possible, almost likely, that you would have research partners with the math skills you desire, who would be interested in collaborating with you precisely because of your computer programming skills, and the ability to automate the creation of research models, and in the process, tackle even more challenging and interesting kinds of problems as a team than either could succeed at individually. Does all the ultra-advanced math have to come from uniquely you? Or is it sufficient to come from a research collaborator, as long as you fully understand it?

So yes, I personally believe you are creating a false dilemma for yourself. I think it would be helpful for you to consider, to the point of writing them out in much more detail than I did above, a RANGE of outcomes that could occur from pursuing this PhD program, and then decide which outcomes constitute "success", for YOU, and which outcomes would constitute "failure". Then, estimate the chances of each outcome coming to pass, 30%, 80%, and so forth. If there are one or more "success" outcomes with likelihood above 50%, then I believe you have your answer. Or, at the least, you are now clear about the question. Do you need "slam dunk" odds of success? or is 50% chance of success good enough for you to get started, confident that you can improve your own odds by hard work and passion?

The other question you musk ask yourself, is "what are the other options", and are any of them as well developed and appealing to you as this one? Because you don't mention them, I can only conclude they are not as well developed or appealing, and that you have put your energies into the option you are asking us about.

The other risk here, is having a well-developed option like you present here, choosing not to pursue it, and then later regretting your lack of action. Picturing yourself five years from now, what will you wish you had done?

Answer 5

If you want to do something in life, you should try it. Try it for a short time defined as less than 5 years.

One of the primary indicators it isn't working out for you is that you'll probably feel great stress if you are barely keeping up with your peers.

Please realize as well that different disciplines focus on different mathematical concepts. For example, if you are interested in becoming a PhD electrical engineer, you probably won't spend any time worrying about number theory or General Relativity. In other words, you don't have to learn ALL mathematics known to mankind in order to succeed in a field that requires skill at some form of advanced mathematics.

Here's something pretty funny for you to absorb. I actually had a professor speak the following paraphrased words in my university philosophy class:

"Most of the research produced at colleges/universities is garbage."

From my own experience at a university I can tell you that there are many dissertations sitting on shelves in back rooms that are doing nothing more than collecting dust. People aren't necessarily as brilliant as you think they are.

Judging by what you've talked about here, I'm laying my money on the notion that you are going to succeed in life.