[BBF Standards] Two separate standards

Mon Mar 17 21:25:54 EDT 2008

Hi Deepak,

I would suggest that in evaluating questions regarding whether a provision
for something should be incorporated into a standard, we use a framework of
what the benefits and costs would be to the various parties involved, rather
than focus on what would be the best way to model the system and reflect all
of its complexities, as these are really separate goals.

For example, I think the primary purpose of publishing standardized
descriptions of "parts" is so that these parts may be used by device
designers.  At a high-level, device designers really need two things when
using a part:

1. A description of what the part does in sufficient detail, legal/technical
restrictions on its use, etc.
2. Sufficient information to allow them to physically incorporate the part
into their device

It's obviously the purpose of the standard to define how exactly to do this.
Incorporating a DNA sequence into the part definition would address much of
#2.

Now consider the composite part, and adding a provision to the standard so
that these parts can described in a manner similar to non-compose parts
(thus addressing #1), but rather than defining these parts by a DNA
sequence, addressing #2 by pointing to the various individual parts which
are part of this composite part, which in turn already define their own DNA
sequences.  What would the consequences be?

- The device designers would be equally happy (but no happier): they still
know exactly what the composite part does because it has its own
description, and their software tools can pull the DNA sequences out of the
sub-parts and combine them together to generate the overall DNA sequence to
synthesize for a device, so it's just as easy for them as before.

- The tool designers are less happy: since the BioBrick Part standard is now
more complex, more work has to go into the software implementation for the
various tools.  This additional complexity is actually the main drawback:
the standard may be no more than a few paragraphs on a wiki now, but these
things have the tendency to grow and grow, so it's essential to avoid
unnecessary complexity until it's needed.

- The designers of composite parts may be slightly happier: They may in
theory have slightly less work to do to publish their part since they don't
have to keep track of the concatenated DNA sequences.  And their work may be
more transparent since you can clearly see what parts they combined from
their part definition.  But I would argue that this shouldn't be the purpose
of a parts standard, and that if one wanted to really understand the
internals of a composite part, they would probably need to read a paper or
additional documentation from the parts' designer anyways.

In this analysis at least it would seem there is little benefit in treating
composite parts specially in the standard unless they are very numerous as
there are real costs but few benefits.  You may know about benefits of
treating composite parts specially that I am missing and would change this
analysis, which would be fine-- I am mainly suggesting that we focus
primarily on the cost/benefits in answering questions like you asked.  This
is one area where the use cases that Ralph suggested would come in handy, as
it would lead to agreement on who/what we are trying to benefit, which
specific proposals could be evaluated against.

Just my two cents,

-Josh 

-----Original Message-----
From: standards-bounces at biobricks.org
[mailto:standards-bounces at biobricks.org] On Behalf Of Deepak Chandran
Sent: Monday, March 17, 2008 12:14 PM
To: Drew Endy
Cc: standards at biobricks.org
Subject: Re: [BBF Standards] Two separate standards

Drew,

Agreed: we do not care about non-functional composite parts.
But is the solution to throw them into the pile of "parts"? You had 
mentioned some weeks ago that the fimH transcription unit (can't seem to 
find that link) is a part, not a device; if I remember correctly, it has 
two promoters/operators that are regulated separately. Such a construct 
seems to match the descriptions of a device, not a part. Is this correct?

--Deepak

Drew Endy wrote:
> Deepak,
>
> One of the great strengths underlying synthetic biology is that we get 
> to invent what we hope to construct (and depict).
>
> We don't have to describe everything that already exists.  We also 
> don't have to describe everything that could be [trying to describe 
> everything is one of the great costs, challenges, and at times, 
> failures of systems biology.
>
> So, in this context, let me ask you a question...
>
> Do we really care about composite parts that are not properly defined 
> devices?
>
> Wouldn't this be like Texas Instruments worrying about every possible 
> combination of transistors, resistors, and capacitors.  Instead TI 
> only considered combinations of parts that resulted in useful devices, 
> and they described only these objects in their Transistor Transistor 
> Logic Data Book.
>
> The only time I've ever seen combinations of parts that are not 
> themselves well defined devices being tracked electronically is during 
> a semi-automated BioBrick standard assembly process.  In this case, 
> the build intermediates need to be tracked during a ligation-based DNA 
> assembly process.  Randy Rettberg developed a software tool to handle 
> this [http://parts.mit.edu/r/parts/assembly/index.cgi]
>
> In any case, remember that we don't have to describe *every* thing, 
> only those things that we care enough to try to make and use.  
> Standards restrict and delimit some possibilities in order that work 
> goes forward and more exciting possibilities can come true.
>
> Cheers!
> D.
>
> On Mar 17, 2008, at 12:38 PM, Deepak Chandran wrote:
>
>> Drew,
>>
>>  This description is clear. The confusion comes when you allow parts 
>> made from other parts to be categorized as "parts" as well -- how can 
>> this be? At the same time, they may not be devices either because 
>> they do not have proper interfaces.
>>
>> --Deepak
>>
>> Drew Endy wrote:
>>> Exactly correct.
>>>
>>> DNA engineers will be concerned with how to construct DNA.
>>>
>>> Parts engineers will be concerned with how to build parts at the  
>>> primary nucleic and amino acid sequence level (and whatever else  
>>> synthetic biological parts get made from).
>>>
>>> Device engineers will be concerned with how parts are best 
>>> organized  to make devices.
>>>
>>> System engineers will be concerned with how devices are best 
>>> organized  to make systems.
>>>
>>> Each type of engineer needs to be able to communicate at least one  
>>> level "down" and "up" in this abstraction hierarchy.  Such  
>>> communication should be limited and well structured.
>>>
>>> Meanwhile, I don't mean to add fuel to any "larger than life T7  
>>> autogene fires," but check out the last 6 pages from the synthetic  
>>> biology comic book, if you've not seen it before.
>>>
>>> http://www.nature.com/nature/comics/syntheticbiologycomic/index.html
>>>
>>> Pages 7-8 detail making a device from parts.
>>>
>>> Pages 9-12 detail some issues of common signal carriers for  
>>> transcription based devices.
>>>
>>> Cheers!
>>> D.
>>>
>>>
>>>
>>>
>>> On Mar 15, 2008, at 2:11 PM, Herbert Sauro wrote:
>>>
>>>
>>>> I think Deepak has made a useful observation. At one level an engineer
>>>> won't care about the DNA sequence, of more interest is the functional
>>>> properties of the boxes he/she is putting together. On the other hand,
>>>> the designer of functional boxes will be concerned with the DNA  
>>>> sequence
>>>> since in designing for example a promoter or gene with a specific
>>>> property the DNA sequence is clearly paramount. It looks like 
>>>> there  may
>>>> be a natural separation here.
>>>>
>>>> Herbert Sauro
>>>>
>>>>
>>>> _______________________________________________
>>>> Standards mailing list
>>>> Standards at biobricks.org
>>>> http://biobricks.org/mailman/listinfo/standards_biobricks.org
>>>>
>>>
>>>
>>> _______________________________________________
>>> Standards mailing list
>>> Standards at biobricks.org
>>> http://biobricks.org/mailman/listinfo/standards_biobricks.org
>>>
>

_______________________________________________
Standards mailing list
Standards at biobricks.org
http://biobricks.org/mailman/listinfo/standards_biobricks.org