NEWS | FEATURES
488 4 FEBRUARY 2022 • VOL 375 ISSUE 6580 science.org SCIENCE
lost hope in him as a scientist at that time.”
UG quietly required Newmaster to fix the
material, Wright and others say. Science also
found plagiarized sections in several of his
published papers, including one on millet
identification in Southeast India (see graphic,
p. 489). Jose Maloles, the paper’s first author,
says it was based on his undergraduate the-
sis, but could not recall how it was drafted.
In a 2020 promotional video made by
Purity-IQ, Newmaster warned about the
risks of data manipulation. “We could have
all the testing in the world,” he said, “and if
that data could be counterfeited or could be
changed in any way it doesn’t really matter
how good the test is.” One month later, in a
training video for the Association of Food
and Drug Officials in which he promoted
Purity-IQ testing, Newmaster displayed
graphics from other sources without credit
and described them as his own work, an
analysis of his talk and PowerPoint slides
shows (see graphic, p. 487).
“Here’s the little experiment that we ran,”
Newmaster says in the video, calling it “a real
life scenario” to guide industry quality con-
trol. But the image he showed, purportedly
representing an analysis of cannabis strains,
is identical to one assembled by other re-
searchers that depicts U.S. arrest data.
Independent scientists identified more
serious problems in Newmaster’s work,
such as an analysis of sarsaparilla—a tropi-
cal plant used to treat joint pain—published
in 2020 with other NHPRA researchers.
Stanford’s Braukmann and Damon Little, a
bioinformatics expert at the New York Bo-
tanical Garden, both examined the genetic
sequences Newmaster provided, and found
those labeled Indian sarsaparilla were actu-
ally near-exact matches for Escherichia coli,
a common experimental bacterium. Prasad
Kesanakurti, corresponding author for the
paper, says the data merely reflected common
E. coli contamination, and offered to provide
the assembled plant sequences for review.
Braukmann says only an examination of the
raw data could clarify what went wrong. The
paper is “an example of poorly done science,”
he says. “It makes me not trust anything that
comes out of [NHPRA].”
THE INQUIRY NOW UNDERWAY at UG was trig-
gered by Thompson, who in 2012 was one
of the first two students to enroll when
Newmaster helped launch UG’s under-
graduate biodiversity major. Newmaster
asked Thompson to work on a paper com-
paring the cost of traditional taxonomic
typing and DNA barcoding for identifying
forest plants. Newmaster provided the sum-
mary data; Thompson had to analyze them
and draft the paper. “We’re getting one-on-
one time with this famous, supersuccess-
ful, important professor,” Thompson recalls
thinking. The resulting 2014 paper in Bio-
diversity and Conservation was his first.
Years later, Thompson grew queasy.
He realized the perfect species identifica-
tion claimed in the paper was virtually
impossible for some of the plants. And
Newmaster had never shown him the raw
data or uploaded it to BOLD or GenBank,
the standard sequence repository. In early
2020, Thompson asked UG to investigate.
“I wasn’t 100% confident that it was fraud-
ulent,” he says. “I was 100% confident that
it was worth asking the question.”
In September and October 2020, in re-
sponse to Thompson’s inquiry, Newmaster’scollaborator Ragupathy deposited thou-
sands of sequence records, purportedly
obtained for the forest paper, in GenBank.
(Around the same time, he uploaded 126
records for the 2013 supplements paper.)
Kuzmina, the CBG scientist, examined the
sequences and found that 80% precisely
matched others submitted earlier for an-
other student’s thesis—collected at a differ-
ent site hundreds of kilometers away.
Thompson—who later also detected some
cases of Newmaster’s apparent image fab-
rication or plagiarism—says UG admin-
istrators slow-walked his request for an
investigation, recast it as an informal query,
and in early 2021 rejected his claims as in-
sufficiently supported. “They thought that
I was just one person, and I didn’t have a
lot of power—that they could squash me,”
he says. He then asked the editor of Bio-diversity and Conservation to conduct his
own review. But the editor deferred to UG.
In May 2021, Thompson self-published
his concerns and posted a related com-
mentary on a popular biodiversity blog,
Eco-Evo Evo-Eco. “Doing this alone behind
the scenes has been incredibly isolating,”
he wrote. “I ... hope that by sharing an
evidence-based critique of our paper some
people will choose to support me.” Indeed,
Hebert soon added a note of support.
Hebert says Thompson’s move re-
vived his own long-running doubts about
Newmaster’s work. He reached out to six
other scholars who could offer authoritative
assessments. They reexamined the forest
paper and also scrutinized the supplements
article and a third paper, published in 2013
in the Canadian Journal of Forest Research,
which found that DNA barcoding of fecal
matter from woodland caribou worked bet-
ter than conventional methods to determine
the animals’ diets. In June 2021, the eight
requested the misconduct investigation by
UG. More recently, some of them also asked
the publishers to retract the supplement
and caribou papers. Hebert says a request
to retract a fourth paper is in preparation.
The allegation letter details the problems
Thompson and Kuzmina detected and many
others. It notes that the papers say barcod-
ing for both the forest and supplement pa-
pers was carried out by the Canadian Centre
for DNA Barcoding, also led by Hebert, but
that the center has no record of that work.
The letter adds that no sequences were de-
posited in BOLD or GenBank before publica-
tion of either paper, and that some of the data
Ragupathy belatedly uploaded in 2020 con-
tradicts the papers’ claims. For example, in
the supplements paper, Newmaster’s group
labeled a product as the laxative Senna al-
exandrina, but the sequence came from a
legume. Moreover, some of the sequences
contained errors that precisely matched
those in sequences previously submitted by
other researchers for several other studies.
In his response to the allegation letter,
obtained by Science, Newmaster strenuously
disputed the concerns. The close correspon-
dence Kuzmina found with samples taken
elsewhere reflected normal species similar-
ity in the forest ecosystems, he said. New-
master insisted his samples were correctly
identified, and that innocent technical errors
could account for matches between rare or
unique mistakes in his sequences and ones
published by other researchers.
Contradicting the papers, he said much
of the barcoding was done not at the Cana-
dian Centre for DNA Barcoding, but at an-
other UG lab, the Advanced Analysis Centre
(AAC) Genomics Facility, or in Newmaster’s
personal “artisanal genomics lab.” Yet he“They thought that ...
I didn’t have a lot of power—
that they could squash me.”
Ken Thompson, Stanford UniversityPHOTO: MACKENZIE URQUHART CRONISH