Time-saving tool checks for conflicts of interest on exam boards

Conflitulus Lattes cross-checks information from the résumés of dozens of candidates and examiners, and may save the organizers hours of work

Gilberto Stam

One of the most time-consuming aspects of organizing competitive entrance exam boards is checking whether there are conflicts of interest between the examiners and the dozens of candidates. The organizers have to manually compare profiles on the Lattes platform, which carries more than 6 million academic résumés and information on lines of research. As part of this task, they look for any coauthorship of articles or projects in common. Generally speaking, researchers from the same area write articles that may have dozens of authors, so the task is painstaking and labor-intensive.

Immunologist Helder Nakaya, a researcher at Hospital Israelita Albert Einstein and professor at the University of São Paulo (USP), was faced with this challenge when invited onto a board to assess 26 candidates.

“To make life easier I developed a code capable of doing the work automatically,” says Nakaya, who uses programming language in his research.

To use the Conflitulus Lattes program, developed by Nakaya and his team, the user must first download the résumés of all candidates and examiners from the Lattes platform, clicking on the XML button to download the compressed file in .zip format.

Compacted or other files may be uploaded onto Conflitulus Lattes after registering.Compacted or other files may be uploaded onto Conflitulus Lattes after registering.

The tool cross-references résumés and conducts initial screening, listing all coauthorships and projects in common between each examiner and candidate, and provides a graph demonstrating possible conflicts.

For a board of 3 examiners and 20 candidates, the full process takes around 10 minutes.

Nakaya developed the program at the Computational Systems Biology Laboratory (CSBL) at the University of São Paulo’s Innovation Center (INOVA USP). Work on the application was supported by Hospital Israelita Albert Einstein, and the São Paulo Research Foundation (FAPESP). “My laboratory develops ‘friendly’ online computer tools, so I decided to transform that into something useful for people,” says Nakaya, the head of CSBL.

Data cross-referencing

“The system uses genome algorithms that convert information such as name and date of birth into DNA sequences, using the letters A, T, C, and G,” explains bioinformatics specialist José Deney Alves de Araújo, CEO of Brazilian startup D2DNA, which specializes in healthcare data analysis, and a former doctoral student of Nakaya who developed the data cross-checking technology.

According to Araújo, the system generates a DNA sequence for each résumé, which is treated as the DNA of a virus or bacteria, with mutations such as errors, deletions, and insertions. An algorithm known as BLAST, used to compare biological sequence data such as amino acids or DNA, identifies similar sequences, even those with errors.

“We align the DNA sequences—in other words the names of the researchers and their publications—to check for any coincidences,” says Araújo.

When the system finds repetitions, it flags up a possible conflict of interest.

Free time

Nakaya published a calculation on his Instagram account of how much time per day is left for a researcher to do science. After teaching, extension, and administration, there was an average of two hours remaining to run experiments, create hypotheses, write articles, and debate results.

“Technology should help us to free up time for the really important things, whether that’s doing science or something pleasurable and personal,” argues Nakaya, Chief Scientific Officer (CSO) of D2DNA. To illustrate this philosophy, the Conflitulus Lattes app estimates savings on hours of work—which already exceeded 24,000 hours, with almost 8,400 résumés cross-referenced to the time this report was published.

Nakaya estimates that cross-referencing a hundred Lattes résumés could save 10 or more hours of work. “Current entrance exams attract more candidates, and people publish more articles, which makes cross-referencing information difficult,” emphasizes Nakaya.

According to the immunologist, the tool has been praised by deans, associate deans, and professors who organize examining boards. Nakaya states that a single user ran the tool 30 times, and cross-referenced almost 1,500 résumés. “We currently have almost 2,000 users registered across Brazil.”

One of the limitations of Conflitulus Lattes is the requirement to download résumés one by one—this is because the Lattes site uses the authentication tool Captcha, which blocks bot access.

Nakaya argues that if the Brazilian National Council for Scientific and Technological Development (CNPq) allowed direct access, then the site could use data direct from its bases, which would make for a quicker, less laborious task for the user. “This would eliminate the need to download the files—you would just need the list of names,” says Nakaya.

“We are looking to integrate Conflitulus Lattes with other academic platforms and databases, which will provide even greater capacity to identify conflicts of interest in different situations,” says Araújo, who believes that Conflitulus Lattes has potential for use in the detection of fraud and data inconsistencies. Araújo says that during a quick check of the Lattes résumé base, he and his team found thousands of typing errors among titles in English. “These inconsistencies can cause a lot of problems, but our technology can get around that in an innovative, efficient way,” he says.

* This article may be republished online under the CC-BY-NC-ND Creative Commons license.
The text must not be edited and the author(s) and source (Science Arena) must be credited.