Unpublished or Perish
a dispute over the intellectual-property rights of a postdoctoral
researcher at RPI, the truth is murky, but the power imbalance
Photo: John Whipple
Roumen Dimitrov cares about science. That much is clear.
Even though the 44-year-old scientist from Bulgaria is facing
the destruction of his career, when he starts to explain
the work he was doing at Rensselaer Polytechnic Institute,
his eyes shine. He stops to draw a picture of strands of
RNA in solution—he was working in “bioinformatics” on software
that predicts RNA conformations and melting curves. The
drawing wouldn’t mean much to a layperson, but Dimitrov
sees in it great possibilities for biotechnology applications.
But then, running his fingers through his graying black
hair, he rapidly turns back to the question at hand. He
has been fired, his H1 professional visa has expired, and
he and his family must return to Bulgaria, where Dimitrov
hasn’t been for 15 years.
The question of why is as complicated as RNA folding. It
hinges on technical matters of how intellectual property
is handled inside an academic lab, but at base it may be
more about the murky territory occupied by postdoctoral
Postdocs are neither students nor faculty, so their roles,
rights and responsibilities in a university are often undefined,
governed more by convention than policy. This often leaves
postdocs at the whim of domineering advisors, but also can
find advisors facing distrustful and frustrated postdocs.
Dimitrov began working in the lab of Michael Zuker, a professor
in the Mathematics Department, in January 2001. He was optimistic
that it would be a good place for him to build up the kinds
of things scientists need to advance a career: published
articles, public seminars and professional connections with
other researchers. “Zuker said, ‘If you want to do science,
we can do it together,’ ” recalls Dimitrov. “I said ‘I want
to have a scientific career, I want articles in good journals,’
and Zuker said ‘I will support you.’ ”
For a year and a half, things seemed to be going according
to plan. Dimitrov was assigned to work on a theoretical
model for RNA folding and melting curves, and to create
software that would allow that model to be tested and used.
This would be added to a suite of programs predicting the
behavior of DNA and RNA that Zuker had been developing with
colleagues and housing on his Web server.
Dimitrov took to the project with gusto, remembers his wife
Dina, who holds a doctorate in microbiology from Uzbekistan.
He “never took a vacation,” she says. “He would work until
3 in the morning.”
But in summer 2002, Dimitrov reached an impasse. The software
he created was giving wonderful results—except, he said,
for a section of code contributed by Zuker. That code, called
“mfold,” made the program fail, or just produce wrong results,
except with very simple cases.
Soon after he told Zuker this, Zuker reassigned the programming
work to his masters students, telling Dimitrov to stop working
on his software because his role was to “help [the students]
when they needed it,” according to Dimitrov.
Dimitrov was stunned. If the masters students rewrote the
software, even if the underlying theory was the same, he
would no longer be considered the author under copyright
law. The masters students’ theses would be based on his
work, but he would be dropped out of the loop before ever
getting credit for what he did, and before any article was
published on the results. “I was used,” he says.
Dimitrov believes that Zuker wanted to rewrite the code
so that he would have control over it. “Once he understood
the only thing that needed changing was his program,” Zuker
became hostile, says Dimitrov. “He wanted to expand the
program and include my part and the students’ [part] so
it’s all considered his.”
Why would Zuker do this? Though getting official credit
and publication was concern number one for Dimitrov, a postdoc,
Zuker had plenty of publication credits under his belt.
But authorship can matter beyond bragging rights or résumés.
the university owns everything created by its employees,
RPI shares the profits of any commercially useful invention
with the inventor. And the program Dimitrov was working
on certainly had commercial applications. During the spring
of 2002, representatives from Integrated DNA Technologies,
a supplier of custom-synthesized DNA to biotech researchers,
visited the lab three times, talking with Dimitrov about
his work. The company was interested in licensing the program,
says Dimitrov, and the inventor would then garner in the
vicinity of 35 percent of the royalties.
Zuker tells a completely different story. He says he reassigned
the programming, reluctantly, because Dimitrov’s programming
wasn’t very good, and he wanted him to focus on the theory,
which was his strong point. “I told him, ‘Your programming
is not up to the standards, [but] you’re quite valuable
to me for your ideas and theories,’ ” says Zuker. “He can
write a program and it can compute, but no one else can
use it. . . . But hey, that’s OK. A postdoc doesn’t have
to be a programmer! We’ve got good students who can do that.”
But if the program wasn’t any good, counters Dimitrov, why
did Zuker represent it as his own in seminars at the University
of Michigan, in order to form collaborations with researchers
there? And why did he previously tell Dimitrov that he was
a good programmer, whose programs were very robust? He also
says there was no reason he should be told to stop working
on his version, even if Zuker wanted to create a more commercialized
or user-friendly version.
Zuker also maintains that Dimitrov’s fears about not getting
credit were unfounded. “We could not have put it all together
if it had not been for things we learned from Roumen,” he
says readily. “There were critical gaps I couldn’t figure
out—he showed me how to do that. I’ll give him credit for
that. I’ll say ‘Yeah, Roumen did this.’ ”
But Zuker also acknowledges that the assigning of credit
for the software itself is a fluid thing. “My student has
written new software to replace everything he’s done. .
. . It’s orders of magnitude better,” he says. “What he
[Dimitrov] should’ve done, he should’ve worked closely with
us and been extremely helpful, [then] he’d get his name
on all the specialized journal [articles]. Instead I have
no intention of putting his name on anything. He blocked
us in any way he could. He gets his name on this one paper
in the Biophysical Journal.”
Dimitrov’s dilemma didn’t come out of the blue. Across the
country, the difficult working conditions of postdocs have
been getting some attention lately. As the number of people
getting doctorate degrees outpaces the number of tenure-track
faculty positions, the number of postdocs has been rising,
and people are spending much longer in these types of positions.
Complaints about everything from pay scales to hostile work
environments to lack of access to university resources have
surfaced in numerous polls of postdocs. In March 2003, a
group of postdocs, mostly leaders of postdoc organizations
at their respective institutions, decided they needed a
coordinated national voice. They formed the National Postdoctoral
Association to advocate for better postdoc rights.
The employee nature of postdocs puts them in a unique situation,
says Alison Reed, the executive director of the NPA. Postdocs
are “not faculty, not a student, not a member of staff,
they’re this nonentity.”
PI [principal investigator, the head of a lab] has the person
as almost a private employee,” adds Avi Spear, a molecular
neuroscientist who was one of NPA’s founders. “There’s often
no grievance procedure, no warning system, no form of complaint.”
As an employee, all the work of a postdoc is under the control
of the advisor, or principal investigator. Faculty say this
is because the postdocs’ salaries are being paid out of
their own grant money.
But some think this exclusive one-on-one power structure
isn’t conducive to good science. “These people are budding,
talented, creative people. They should not be held at the
whim of one person who has often little management training,”
says Spear. Although Spear says he doesn’t have enough details
to comment on Dimitrov’s case directly, he does say, “There
are people who steal your ideas and steal your credit. The
postdoc has very little leg to stand on if the PI wants
to do that.”
The situation is even worse for foreign postdocs, whose
visa status is dependent on their employment, and therefore
on their advisor. They can “play with me because I need
a green card,” says Dimitrov. The number of postdocs coming
from other countries has increased dramatically over the
past 15 years. According to a National Science Foundation
study, the number of foreign postdocs in science and engineering
rose by 8,000 from 1988 to 2000, while the number of U.S.-citizen
or permanent-resident postdocs rose by only about 1,500.
As for forcing a postdoc to turn over his project for grad
students to work on, Spear says, “It’s mean, it’s abusive,
but it’s allowed. What you work on in a lab is the property
of the lab.” But, he adds, “It’s not common practice.”
Dimitrov thinks it’s too common. “It’s a system problem,”
he says. “It smells as a business, it doesn’t smell as science,
doesn’t smell as equal opportunity.”
are considered cheap labor without any inherent rights,”
agrees one RPI postdoc who doesn’t wish to be identified.
“[In the United States] postdocs are [often] taken to do
grad-student work. No one ever did that for me. I had to
do it on my own.”
Age discrimination also rears its head from time to time,
and was one of the nails in the coffin of Dimitrov’s trust
for Zuker. When he was reassigning the programming, says
Dimitrov, Zuker told him he was too old for a career in
science and should resign himself to being a postdoc assisting
other professors. Zuker vehemently denies that he ever made
any such comments.
But others have seen that sentiment expressed among some
American faculty. “I think that they think that 39 or 40
is the [limit],” says the unnamed RPI postdoc. “I don’t
know why. I was always used to that people were rewarded
for their experience and length of time. [But] people without
experience are younger and they’re cheaper.”
Within RPI, there is an atmosphere of fear among the postdocs,
especially the foreign ones, says Dimitrov.
The Dimitrovs say they know of several international postdocs
who were forced out of their positions after they had turned
data over to their advisors, but before they could publish
the results and receive the credit. Those postdocs, one
from Bulgaria, one from Brazil, one from Colombia, could
not be reached for comment since they had to return to their
home countries once their employment ended—and RPI did not
keep contact information for them.
RPI administrators say such things are not happening at
RPI. “I’ve never heard of that happening at a place like
this,” says Charles Carletta, RPI’s general counsel. “The
academic community at research institutions—in the U.S.
anyway—have a lot of integrity and are usually very careful
about attribution amongst themselves.”
But two RPI researchers, who want to remain anonymous, say
they believe Dimitrov. One actually warned him to get out
while he could. Another said both credit stealing and age
discrimination are rampant in the United States, and Dimitrov’s
was not the only case that had arisen at RPI.
Such an atmosphere could be enough on its own to make someone
in Dimitrov’s position distrustful. But the conflict with
Zuker was not Dimitrov’s first tangle with the system. He
had originally been hired to work in the lab of Chris Bystroff,
an assistant professor of biology. Dimitrov says that relationship
was strained from the beginning. “The professor wanted complete
control, just someone to produce results for him,” says
Dimitrov. He said he was constrained from doing anything
independently, and told, “If you come here [to this country]
and want to stay, you have to work for me and give me everything.”
Dimitrov says he was assigned to work on small problems
and gaps in Bystroff’s research, and when he asked if he
would get credit for solving them, he was told emphatically
no. “I need your results and what I do with it is my business,”
Dimitrov says Bystroff told him.
Bystroff remembers it differently. “He worked for 10 months,
he couldn’t account for his time working in the lab,” he
recalls. “When I pressured him for some results, he was
reluctant. I assigned him some projects, he didn’t work
on them. He never gave me any kind of a scientific result.
. . . [But] there was never a question that he’d get credit
for what he’d produced.”
Dimitrov insists he did everything Bystroff asked of him,
quickly ticking off a list of projects as examples. Dimitrov
says he asked for a transfer; Bystroff says he asked him
After their first falling out in Summer 2002, Zuker’s and
Dimitrov’s next battle was over publishing an article about
the new theoretical model. Dimitrov, feeling credit for
the software slipping away from him, was determined to at
least have one official publication about the theoretical
model under his belt.
He found some limited experimental data for which the programs
as they stood returned good results, and wrote up an article.
Zuker said he didn’t want it to be published until the code
was fixed to work more consistently for a wider range of
After much argument, however, the paper was submitted to
the Biophysical Journal on Dec. 4, 2002—with Dimitrov’s
name first and Zuker as a correspondence author. It is common
practice in a lab situation to add an advisor’s name to
anything produced within the lab. The paper was well-received
but sent back to have some deviations explained.
Those deviations also stemmed from Zuker’s mfold program,
says Dimitrov, and since Zuker wouldn’t explain that code
to him, he eventually sat down and wrote an algorithm that
replaced the mfold program. Though it was still limited
to a specific type of case, the new algorithm worked perfectly,
Dimitrov says, for much longer strands of RNA than mfold.
“I used my programs on very long sequences from the company
[IDT], and it hit them perfectly,” he enthuses.
He was very excited, but since he had been told to stop
working on the program, and it appeared that Zuker didn’t
want the article published until the code would be entirely
under his own copyright, Dimitrov didn’t share the algorithm
with Zuker. He rewrote the article with the new results,
but didn’t include the actual code there either.
The problem was, as an employee of the lab, Dimitrov was
required to share his work with the “principal scientist.”
Dimitrov knew this, but says at that point he was just trying
to do what he could to salvage his career. “I was doing
this to survive,” he says. “All I wanted was evidence I’d
Dimitrov sought the advice of the intellectual-property
office at RPI. “They told me if you have this problem in
your lab, you should [apply for a patent]. This was RPI,
not me,” insists Dimitrov. “You have copyright on the code,
but they can rewrite it so its not yours,” he recalls being
told by Paul Fredette, assistant director of what is now
called the Office of Technology Commercialization, “so you
need to patent the whole model.” Fredette says he does not
remember the specifics of his conversation with Dimitrov.
The office filed a pre-patent form for Dimitrov in April
Under U.S. patent law, who contributed what to an invention
is determined by patent officials. Distinctions of faculty,
postdoc or student do not come into play. The patent application
made Dimitrov even more determined to keep his algorithm
to himself, so he could prove to the patent officials that
he was in fact its creator. RPI, however, does own the patent
rights on anything made with “significant use of RPI resources,”
which Dimitrov acknowledges. He just wanted credit.
Meanwhile, the new version of the article was accepted by
the Biophysical Journal in May 2003. It has not yet
appeared in print. Zuker has requested extensions, missed
deadlines for returning proofs, and, more recently, removed
certain portions of Dimitrov’s additions, because, he says,
he doesn’t want a publication to go out with information
he cannot replicate. In October, the journal’s Web site
claimed the article would be included in November’s issue.
As of this week, mention of it has been removed from the
Dimitrov fears the piece will never see the light. In fact,
when he chose to cancel a settlement with RPI this October—which
he said he’d signed under intense pressure—Dimitrov says
that human-resources vice president Curtis Powell stated,
“Now the article will never be published.”
The struggle over the unrevealed code peaked this summer.
In July 2003, Zuker wrote a letter to Dimitrov, complaining
about Dimitrov’s secrecy and his reluctance to work as part
of a team, and sent copies of it to six administrators.
“It seems to me that your secretive behavior comes from
a false belief that you have sole intellectual-property
rights to what we have developed,” the letter ends.
Finally, the chairman of the Mathematics Department, Donald
Drew, brokered an agreement. Zuker and Drew promised Dimitrov,
in writing but without the approval of human resources,
a three-year contract in return for Dimitrov’s agreement
to direct Zuker’s masters students, so that their software
could reproduce the results Dimitrov was getting with his.
Feeling backed into a corner, Dimitrov agreed.
Dimitrov began to work with one of the computer-science
students, trying to give him the theoretical background
to understand the model he was programming for. The student,
however, said in an e-mail message to Zuker that Dimitrov
was refusing to give him his code. Dimitrov pointed out
that that was not what he had agreed to do.
On Aug. 27, Zuker and Drew showed up at Dimitrov’s office
to tell him he was fired. Though they gave him the required
30-days notice, his termination letter required him to vacate
the campus that day. In fact, he was escorted from the building
by a security guard, who told him he was “persona non grata”
at the institution from now on.
Zuker doesn’t dispute this. “Why was it done in such an
abrupt manner? Why was he escorted out of the building?”
he says. “I wanted to look at his computer without warning
[him] so I could see the files, which was his right and
his duty to expose to me.”
Dimitrov fought his firing, appealing to RPI President Shirley
Jackson for help. Jackson’s office referred the matter to
Powell of human resources, who promised, in writing, an
investigation by “external legal counsel.” RPI hired Pattison,
Sampson, Ginsburg & Griffin, its own attorneys, to carry
out the investigation. That’s an odd definition of “external
counsel,” says Christina Diamante, a friend of the Dimitrovs
who attended most of their meetings with Powell.
don’t use words like ‘external’ and ‘internal,’ ” says Charles
Carletta, RPI’s general counsel. “They requested an investigation
and we went to the firm that does our investigations.”
The investigators concluded that Dimitrov’s termination
was warranted, saying he had demonstrated “a failure to
adhere to Rensselaer’s policy regarding intellectual property.”
This wasn’t surprising to postdoc advocate Avi Spear. “Whenever
there’s a dispute between a PI [principal investigator]
and a postdoc, the postdoc always gets sacked,” says Spear.
“Whoever’s at fault is variable, but the net result is always
the postdoc gets fired.”
According to Diamante, when asked about their results regarding
the charge of age discrimination, Powell said the Dimitrovs
had never brought the issue up. She denies this strenuously.
“I heard them mention it,” she says. “I have it in my notes.”
The issue was also raised in Dimitrov’s letter to Jackson,
which prompted the investigation.
While he was contesting his firing, Dimitrov had also begun
to look for other work, in order to be able to stay in the
country. People with H1 professional visas have one month
after the end of a contract to get another job or their
visa expires. He had plenty of nibbles, but they all ended
in the same suspicious way. Despite a promise from Powell
to “erase all black marks,” employers would be initially
interested, and then suddenly turn him down after calling
RPI for references.
Powell refused to comment for this article.
In the midst of all this, Dimi- trov insists that his gripe
is not personal. His message, he says, is that there is
a systemic problem—universities, particularly through the
institution of postdocs, are becoming more and more like
private companies, and less conducive to good science.
Powell told me I have to do what my manager is telling
me,” says Dimitrov. “Is this a university environment, or
a private company where I’m not doing science, but moving
objects from here to there? Postdocs in this country have
no rights. We work and feel as horses, not scientists.
are just a tool, they just use you,” he adds. “I don’t like
[that] they will do this to other people.”
The NPA’s approach to these problems is to advocate for
formalized, and standardized, agreements made—in writing—between
the postdoc and the advisor before they begin work. These
agreements should include the postdoc’s scope of work and
how credit will be assigned, as well as expectations regarding
use of institutional resources, mentoring and progress evaluations,
says Reed. Whether such an agreement would have prevented
the problems between Zuker and Dimitrov is unclear, but
it might at least have made the rules of engagement plain.
Meanwhile, the Dimitrovs are preparing to leave. They have
sold their belongings to purchase airfare back to Bulgaria,
and jumped through dozens of hoops to secure visas for Dina
and their 6-year-old daughter, who was not born in Bulgaria.
Because they had prepaid their daughter’s tuition for the
fall, they have no savings. Their friend Diamante has set
up a fund to help ease their transition, because employment
prospects are likely to be dim upon their return. Since
Dimitrov has been gone, the country has transitioned to
capitalism, new social systems have sprung up and most professors
he used to work with have retired. “I am outside of the
system,” he says. “Nobody knows me. I have to start from
came here in hope,” says Dina sadly, “but in the end, they
closed all the doors.”