Sunday, February 10, 2019

A Moral Basis for Ethical Genetic Engineering

I recently read a news report saying that two colleagues of mine have been trying to put together a business to create genetically modified human babies with added "good" traits. My immediate reaction to reading this news was deep upset and strong moral distaste.

But why? If I believe that some types of genetic engineering are wrong, while other types are permissible, what is the actual basis for my personal moral judgement? Relatedly, what does that say about the code of ethics that I would want practitioners of the field to follow? (For purposes of this discussion, I will use "morals" to refer to personal evaluations of right and wrong, and "ethics" to refer to the practices a community uses to try to avoid bad moral consequences).

I've been mulling these questions over personally for several years now, driven both by my own thoughts and my conversations with friends, family, and colleagues. More recently, I've been starting to have these conversations with other people at BBN as well, as our synthetic biology group grows and we consider an ever broader set of possible opportunities to pursue. Which calls for proposals should we embrace, and which should we pass over because we do not approve of their direction?

I think these are extremely important questions to think about carefully and have a clear understanding of where one's judgements are actually rooted. On the one hand, our minds easily conflate "unfamiliar" with "wrong", and history is full of lessons on how words like "unnatural", "improper", and "distasteful" have simply been codes for prejudice that has had to be unlearned one small step at a time. On the other hand, history is also full of lessons about how easy it is to step into morally abhorrent positions and actions one seemingly reasonable step at a time.  Having a clear understanding of the basis of one's judgements is an important defense against both of these failure modes.

A Few Assertions on Morality

For a starting point then, let me begin by leaving any potential deities out of the discussion. Instead, let me start with a few grounding assertions that I think most will find non-controversial:

  1. Conscious minds are precious. I know I treasure my existence, and expect that most others generally do as well. My circle of empathy extends at least as far as nearly all living humans and a lot of the more brainy animals.
  2. I should treat others as I would like to be treated.
  3. Deriving from the first two: a person's autonomy of choice should be respected, at least so far as it does not infringe on others.
  4. It is better to avoid unnecessary suffering. Sometimes suffering is necessary or unavoidable, but given a choice it is generally preferable to have less suffering in the world.
  5. We often make mistakes. This is especially true when dealing with new or poorly understood things and large-scale or long-term consequences.

These statements are by no means the whole of my moral system, and there's lots of grey areas to explore with regards to their definitions, boundaries, and conflicts. They are however, some good basic guardrails for my thinking: anything that clearly starts to violate one of these assertions is a place where I don't want to go.

Moral Judgements on Genetic Engineering

So let's start looking at genetic engineering as a subject of these moral judgements.

First off, is there anything about the creation and editing of DNA (or similar) that is inherently morally problematic?

The necessary materials and equipment involved are relatively cheap and easy to obtain from normal sources. So with respect to the material resources involved in genetic engineering, we're talking about moral issues akin to those involved in eating cereal or buying clothing, not anything specific to genetic engineering.

Likewise, I see no inherent problem in modifying the DNA of living creatures. There are examples that I find clearly in support of my moral values, such as the development of gene therapy to correct otherwise fatal or debilitating genetic diseases.

It seems then that any moral judgements that I make are not grounding in the technology itself, but in the bad effects resulting from choices that we may make about how to use it. In short, genetic engineering poses moral and ethical challenges because it is a disruptive technology that gives us choices that we did not have before.

What sort of bad consequences am I concerned may result from poor choices in the use of genetic engineering?  Well, some are just the usual concerns related to the potential for any disruptive technology to reorganize money and power in societies, but those are not specific to genetic engineering. When thinking about the specific technologies, here are some key things that I would like to avoid that I also think should be relatively non-controversial:

  • Injuring or killing people: obviously counter to the moral values I've expressed above
  • Degrading people's autonomy: likewise, obviously counter to moral values.
  • Damage or destruction of infrastructure: creates disruptions that tend to involve suffering, injury, and death.
  • Disruption of ecosystems: another source of disruption, and often unpredictably so
  • Splitting humanity into more than one species: we have done badly enough morally with "other" groups when we are all at least members of the same rather homogeneous species.
  • Significant loss of human diversity: seems likely to involve degradation of autonomy and to lead to increased fragility.

Again, this is by no means attempting to be a complete list, but is at least a good set of guardrails to begin with. If a project has the potential to make one or more of these scenarios more likely, then that is clearly a moral hazard to concern ourselves with.

Ethical Genetic Engineering

Turning from morality to ethics, how do I think that concern about potential consequences should affect our actions? First and foremost, I am strongly aware of the fact that we humans frequently make mistakes and that some consequences of genetic engineering, once set in motion, might be quite large-scale in impact and also quite hard to reverse. This leads me to embrace a version of the precautionary principle: whenever considering a research or development choice that may have a major moral impact, I would hold that one should move slowly and incrementally, step by step building up knowledge, precise and predictive models, and increasing levels of consensus regarding the morality of particular choices and their consequences.

In my view, then, an ethical approach to decision-making in genetic engineering ultimately boils down to a relatively simple core:

  • For any potential project or technology, one must assess the degree to which it increases risk on "the checklist of bad consequences."
  • The closer one is to real-world applications, the more predictive certainty is needed in this risk assessment.

Great complexities, of course, may arise in actually making these evaluations, and that is the point: if you want to take risks with lives, species, or ecosystems, you'd better be able to establish with great depth and certainty that the risks you want to take are truly low.

Returning to the Matter at Hand

With this enunciation of my approach to the ethics of genetic engineering, I think the reasons for my reaction to the news I read becomes quite clear. First, the report spoke of plans that could take potentially grave risks with thousands of human lives within a mere few years (e.g., what if a "good" modification turns out to have nasty side-effects on the children bearing it a few years down the line). Moreover, in the report there seemed to be no evidence that those proposing those plans were even thinking about the risks, let alone making reasonable plans to assess and mitigate these risks. Perhaps the report is wrong, and perhaps those colleagues will communicate facts to me that would cause me to change my judgement of their work. For now, however, the facts that I have been made aware of certainly seem to show a serious violation of the ethical principles that I espouse.

Making this assessment has been quite a challenge, and I expect that I will revisit it over time to see if there are things I want to add or to adjust. For now, however, I am satisfied to rest with this increase in understanding of where I stand and why on the moral and ethical questions that are involved in genetic engineering.

In short: "First, do no harm."

Thursday, January 17, 2019

The end of an era

One week from today will be the official end of an era for me. After nearly 23 years, more than half of my life, I will no longer have an account with MIT.

My MIT email address was my first "real" email address, at least in the sense that I can no longer remember definitely just what my high school email addresses actually were: you'd have to dig into the prehistory of AOL or old shareware repositories to find that information. I signed onto Project Athena in August of 1996, was indoctrinated into the joys of "mh", and began my tumultuous undergraduate career.

I stayed at MIT for 12 years as a full-time member of its community: four years as an undergraduate, one as a Masters student, one in a strange superposed state of both Masters and Ph.D. programs, which mightily confused the registrar's systems, five more years of purely Ph.D., and then a year of transitional postdoc while I figured out what to do, ultimately departing for my current employer of BBN in 2008.

Self-portrait as a postdoc, in my old office in Project MAC at MIT CSAIL
Even after leaving, however, I maintained my affiliation and strong collaborations. In the first couple of years, I was actually also still running the same MIT projects that I had been running while I was still actually employed there, and so of course I was always on campus to meet the students who were working for me. Other collaborations started up thereafter, and one way or other, it tended to be the case that I was working on campus at MIT at least half a day in every week. Keeping research affiliate status for me made a lot of sense for all involved.

Back in those early post-departure days, I used to still be much more involved as an alum in student group activities as well.  One of my long-running joys, which I still miss, was running live action role playing games with the MIT Assassin's Guild. More regularly, however, I also continued to be a volunteer librarian with the MIT Science Fiction Society, and every week would spend two hours as the on-duty librarian holding open the worlds largest publicly browsable collection of science fiction. I needed my card and my affiliation to be effective at those duties as well, and I enjoyed them much: the Assassin's Guild as a heated activity of creative passion and adrenaline, MITSFS as a cool oasis of two calm hours of mostly only reading.

When I followed my wife to Iowa in 2013, however, the actual "showing up on campus" part stopped happening. I resigned as a librarian, and I'd already mostly stopped writing and playing games, as new parenthood and professional travel began to squeeze that time more and more.  My collaborations have continued, but with me no longer actively on campus or needing special access to resources, there's not as much point in having me still maintain an active affiliation.

Sometime last summer, my affiliation failed to renew, and I didn't notice. When I got my account deactivation warning, I pinged the collaborator who'd been sponsoring me, but neither of us got around to following up. And really, the fact that it just wasn't making my triage list as "important" any more was the sign that it was time to let go.  I'm no longer an active alum, and I don't need a research affiliate status to be an effective remote collaborator, after all.

And so, over the past two weeks I've been packing up to go electronically. I've redirected my non-BBN mirror of my professional webpage away from MIT and over to GitHub. I've copied over all of the material from my old Athena account (finding and revisiting some amazing old memorabilia in the process). I've even gone through every email received at the old account in the last year and switched over all the ones I cared about. I'm as ready as I can be to let go.

Goodbye old friend, old email address. I never like to truly let anything be gone, but I'm not there any more, and at least I've still got my alum account.

Saturday, January 12, 2019

Taming emergent engineering

Understanding and engineering emergent behaviors is one of the long-standing challenges of complex systems.  Over the past fifteen years, one step at a time my collaborators and I have been pinning down the engineering of emergent behaviors.  Our most recent publication, however, represents quite a major step in the project.

"A Higher-Order Calculus of Computational Fields", out this week in ACM Transactions on Computational Logic, finally puts a solid mathematical link between collective phenomena and local actions.  In this paper, we present not one but two equivalent semantics for aggregate programs: one in terms of local actions of devices and the other in terms of collectives extending across space and time.  Every field calculus program expressed in one view can be automatically translated to the other, from global to local and from local to global.  We've been working with this result informally for many years, but now we have rock-solid mathematical proof.

Now combine that with "Space-Time Universality of Field Calculus", a paper we published last year demonstrating that every computable function over space and time can be implemented using field calculus.  That tells us that, no matter what emergent behavior you might be dealing with, if it is physically possible, there is guaranteed to be a program that can be expressed in our simple language that can both describe the collective behavior and be applied to produce it from local interactions.

This doesn't mean we can predict the behavior of any old system out there.  Just because you know there is a description doesn't mean it will be easy to find it, or that said description will be simple. Likewise, it might be difficult to understand the implications of a program. But having a simple language that is guaranteed to cover all of the relationships of interest can make a very big difference in just how hard that search space is to navigate.

Unfortunately, I don't really recommend that you read either paper unless you love wading through heavy mathematical symbology.  Ultimately, once you wrap your head around the mathematics, the core ideas of each paper are fairly simple and elegant, but there's a lot of supporting details that have to be dealt with, systematized, and pinned down with mathematical variable names.

Next step: making a more digestible summary of the key results available to the wider community who may be interested.

Example of resolving an aggregate function call over space and time in higher-order field calculus.

Monday, September 17, 2018

Ten years at (Raytheon) BBN

Ten years ago, I started my job at BBN Technologies, my first professional step outside of my graduate school environment.

Much has changed since then: the focus of my research has shifted; the networks of colleagues I work with have grown and changed; BBN was swallowed up by Raytheon; I got married, moved to Iowa, and had two kids. I am more comfortable and confident in my skills and my abilities, I've learned to manage both my arrogance and my imposter syndrome better---and also learned that those are just two sides of the same problematic coin for me, and are likely chronic challenges that will not go away. I've learned how to more effectively say "yes" and I'm still learning to say "no."

Tiniest Moose, helping me with my work on a business trip to California.

At its core, however, my world of research and professional life is much the same. I wake up every morning entangled in the delicate balance of work that might have a profound impact on our society and work that will be completely irrelevant before it is completed (and sometimes little way to tell these apart). I take joy in my collaborators and the artifacts we produce, the satisfaction of programs working and data-points that form a beautiful line, the hope and anguish of proposals and papers submitted, rejected, and accepted. My day is a day of the craft-work of the scientific, in all of its prosaic glory, and I have every anticipation that I will find it no less engaging years from now, even if someday I end up somewhere else in terms of my career.

For its part, Raytheon, in its infinite wisdom, has informed me that in honor of my ten years of service, I am to be awarded a gift picked from a menu of some intriguingly "safe" and mediocre choices, like a fancy dart board, designer shades, a bike rack, a glowing bluetooth speaker, or a package of Omaha steaks. We picked the carpet cleaner. Happy anniversary!

Monday, September 10, 2018

What's your bus factor?

My wife and I had our second child just under three weeks ago, and as I'm slowly beginning to find my equilibrium in the midst of my parental leave, I find myself contemplating my personal bus factor.

The bus factor is a tongue-in-cheek name for a measurement of a project or organization's level of robustness. The proposition is this: let's say that some critical people in your organization are out to lunch one day, and while crossing the street, they get run over by a bus. The bus factor is the minimum number of people who, if they go under that bus, will result in the project or organization being badly disrupted. It's a wonderful and horrible thought experiment that asks us to face the question: life happens---whether it be a bus or a baby, marriage or divorce, cancer or a long overdue vacation---and how does that affect the world of work?

Obviously, as an organization you don't want a low bus factor. If your project has a magic guru without which all is lost, then your bus factor is 1 and sooner or later all will be lost. So from the perspective of resilience, a higher bus factor is always better.

As an employee, however, a high bus factor is actually a very bad sign for your value to the organization. If your organization has (or thinks they have) lots of people just like you, then they probably won't be valuing you as much as you would like them to. As a researcher, my value is ultimately in my expertise and my ability to deploy that in ways few others can, so my bus factor had better be fairly low. So it would appear that from the perspective of the individual employee, a lower bus factor is always better.

When your bus factor is too low, however, that's bad for work-life balance. If you're critical path on everything you do, then the ups and downs of projects can't be shared with others. Every crisis is your crisis, and if you take time out then things break and you let everybody down. At its worst, your bus factor is effectively less than one and you're always doing overtime just to stay afloat.

So, what's your bus factor?

In my own personal transition over the past year and a half, as we've grown the synthetic biology team at BBN while still continuing to execute on projects that started well before, my bus factor has definitely been under one at times. Preparing for paternity leave, however, became a very interesting exercise in evaluating how things had changed and where I needed to rethink how I had things organized at work. I started well ahead of time, listing out all of my different responsibilities and seeing what were the truly critical things that needed me to do them before my gap, then working with colleagues to plan out how to cross the time of my absence. Since you never know when a baby will come, I started warning people about my likely disappearance weeks ahead of time (and there are wonderful labor prediction tools now available online to make it quantitative!). My bus factor in prediction seems to be about 1.5, meaning some things will break if I am gone too long, but my team can probably run for quite a while without me, given our preparations and the competence of the people that I collaborate with.

And now I'm gone. I've got a lovely baby girl, an older daughter who seems to be adapting well, and our sleep schedules are as unpredictable and in flux as any parents of a newborn might expect. I've been quite solidly off my email and not taking calls. I'm grateful for the privilege to have this time off and spend these early days at home, and think it's a tragedy that here in America so many who work are unable to take such time. Babies and parents both deserve better, and our economy could most certainly afford it if we had the will as a society.

When I come back, I will find out how right or wrong we were about my bus factor and our preparations. Maybe at this very moment a project is going down in flames and I will have to deal with terrible things in the moment when I first connect and read what I've been missing out on. But I hope not, and am grateful for the professional community that has let me continue to try to walk this work-life tightrope as I balance.

Thursday, June 21, 2018

Big paper out today: units matter in biology!

This is a big one: our paper out today, "Quantification of bacterial fluorescence using independent calibrants," is the official peer-reviewed presentation of the results from the 2016 iGEM interlaboratory study.  After spending a year or so digesting all of the data for publication, the bottom line is this: everybody can and should calibrate their fluorescence measurements.

Here's the key figures of the paper, showing just how much error reduced when using an independent calibrant to put units on your measurements. And notice those orange bars in the middle: that's how well you can do with relative units based on a control strain of cells. It's better than nothing, but still far worse than with an actual independent calibrant, because there are so many ways your control strain can get messed up in the same way that your experimental strains are getting messed up.

It's cheap. It's easy. High school students and undergraduates can do it.  And so should every other biological researcher or engineer measuring cellular fluorescence, especially those working in synthetic biology.

Sunday, April 08, 2018

Linking biological designs and experimental data

One of the biggest points of friction in my professional life is the disconnect between the design of an experiment and the data that comes out of it. Not in any deep or scientific sense, but in a boringly practical sense of "How do I know what's in file MyRun_F05_039_pXK405.fcs?"

When I'm working with experimentalists and analyzing the data that they've produced, in order to make this connection, I get sent spreadsheets with colored cells and personal shorthands, or unintentionally cryptic emails, or scans of tables with hand-written notes. Then I make my best guess as to what's being encoded there and start organizing file names into scripts to run my analysis. The actual process of analysis is often very fast, only a few minutes, but for a good-sized experiment it can take hours to set it up to be able to run.
Example of fairly typical current integration of biological data with experimental design.
Even then, our pain isn't over, because there's a major challenge in comparing across data sets, especially when working with multiple people on a project or across a project spanning many months or even years.  Is the control the same as it was two months ago? What does "same" even mean, exactly? I had a data-set go completely wonky once because the experimentalist working with me had run out of one plasmid and substituted another that they thought should be equivalent but had an extra "unimportant" gene on it.  The descriptions that I got gave the same descriptor to refer to the new plasmid as they used for the old one, because of course they were only describing the "important" parts of the construct. We lost at least a month of time on the project.

All of this can be simplified if we get automated software tooling involved, so that with minimal human involvement we can link data to laboratory samples, samples to the descriptions of what they are supposed to contain, and designs for DNA to the biological functions and interactions that they are intended to produces.  For that to work, we need to agree on how we are going to describe those relationships, and thus I believe that the most critical part of what our newest release of the Synthetic Biology Open Language (SBOL), version 2.2, gives to us, along with some tools for describing combinatorial designs.  Version 2.2 has just been officially published as a free journal article, and we're well into putting these new linkages to use in several programs, as well as organizing a workshop to teach people how to link these and other tools together

Step by step, we are getting closer to removing this persistent source of friction and error in our biological studies.

Sunday, March 18, 2018

Diagrams showing structure and function in biological organism engineering

We've just had official publication of another major step forward in turning synthetic biology into a well-organized field of engineering: the SBOL Visual 2.0 standard. This is a big one, because it means we have a clear way not only of summarizing genetic structure (as we have had since SBOL Visual 1.0), but also of showing the interactions of genes with proteins and other molecules in order to actually affect cellular functions.
Example of an SBOL Visual 2.0 diagram, showing a system with two functional units: one producing the regulatory protein TetR, which in turn represses the other's production of green fluorescent protein (GFP).
Everybody's been drawing diagram sort of like this already, in the papers that they publish, but there hasn't been any agreement on how to do so, and so every diagram's a little (or a lot) different, with no good way to make sure that you really know what somebody's diagram means besides reading the whole text in detail---and sometimes not even then. Now, with this standard, we have such a system, and we just need to work with folks to keep spreading the word so that people are aware and can understand how following the suggested guidelines will help them by making it easier for others to read what they have written.

Friday, March 16, 2018

Good Measurement Practices

As we work to promote awareness and use of good scientific measurement practices in iGEM (the International Genetically Engineered Machines competition), we've just posted an educational video with me giving a (hopefully accessible) introduction to four simple principles of good measurement practices.

Tuesday, February 06, 2018

The LOLCAT Method

You probably think the title of this post is a joke. Well, it is, but probably not in the way that you think it is.
LOLCAT helping me with SCIENCE!
You see, back in the waning days of my grad student career, I started working with an ambitious and enthusiastic young undergrad named Sagar Indurkhya who wanted to work on better ways to design synthetic biology circuits. I was just getting into the area myself, and our efforts quickly wandered sideways, from work on circuit design to work on simulators. Sagar was using stochastic simulators and found (as many people do) that they were way too slow for his taste. So he went to town on the optimization problem, finding all sorts of crazy ways to improve the speed, from highly general (factoring reactions to improve scaling properties), to super-specialized (making his own specialized virtual machine). Happy with the remarkable improvements in speed that we'd gotten, we decided to write it up and, liking publications without paywalls and having no particular reason to send it anywhere else, we sent it to PLOS ONE.

In the process of writing things up, however, we needed to give the algorithm a name, and one fateful day Sagar asked me: "Can I name it anything?" I said sure, and he continued, "Even something silly, like LOLCAT?" I hesitated, but couldn't really find any particularly good argument against it besides the fact that it was silly, which at the time didn't seem to me to be a sufficient argument against. And if it was a problem, the reviewers would ask us to change it, right?

Not a peep. I just looked back through and found that the reviewers were perfectly happy with our absurd title, engaged seriously with the paper to provide a sound and sober analysis of the LOLCAT method that resulted in significant improvement in manuscript presentation, and then the paper went through for publication. And then I mostly just forgot about it.  I don't use stochastic simulations very often, and when I have it's typically been on much smaller systems, so I just haven't ever had reason to use the work myself.

But others have. I was reminded of the paper this morning, in fact, by a citation alert. After a long period of dormancy, the LOLCAT method is gathering citations as reaction network simulations grow and people are apparently finding it to be of significance in their work. As of this writing, it has received 18 citations---not huge, but definitely showing a significant impact.  I am profoundly ambivalent about this fact: happy that it's a useful piece of work, cringingly embarrassed at my early career naiveté, yet also defiantly proud of our little joke. We didn't even have the good grace to try to make the name an acronym.

It's out there still, and will be in the scientific record forever after, for good or ill: "Reaction Factoring and Bipartite Update Graphs Accelerate the Gillespie Algorithm for Large-Scale Biochemical Systems."  The LOLCAT method.