Blood lies
An African Australian’s DNA nightmare by Oscar Schwartz
When the Department of Immigration refused a partner visa for Daniel Tadese’s Ethiopian wife on the basis that a DNA test showed they were related, the couple knew a mistake had been made. Their ongoing ordeal demonstrates that Australian authorities rely on faulty science that is especially prejudicial against Africans.
On February 24, 2012, Daniel Tadese and Genet Abebe were married in Addis Ababa, Ethiopia, after having been introduced around one year earlier by a mutual friend. Tadese was living in Melbourne and Abebe in Cairo, Egypt, so their relationship started out long distance. After getting to know each other over the phone they planned to meet face to face in Ethiopia’s sprawling capital, where Abebe’s family lived. There, they spent almost every day together for three months, deciding, as Tadese’s visit came to an end, that they would get married before he returned to Australia. Three days after their wedding, a grand occasion of around 300 people, their respective families held a kleklek, a traditional gathering for the in-laws to mingle. The Tadese and Abebe families talked, drank, ate and bid the newlyweds good luck for their new lives together.
Tadese left for Melbourne and Abebe travelled to the Australian High Commission in Nairobi, Kenya, to submit an application for a partner visa, so she could join him in Australia, where he was a citizen. Shortly after, she told Tadese that she had received a troubling note from a delegate from the immigration department. The officer she then met had apparently advised the department that there was a resemblance between Tadese and Abebe. Not long after, both Tadese and Abebe received letters requesting that they undertake DNA testing to support their visa claim. Somewhat perplexed, they decided that it would be best to comply with whatever the department was asking. “I was happy to do it because I had nothing to hide,” Tadese tells me. “I wanted to be back with my wife.”
DNA testing has been routinely used in immigration cases for at least two decades in Australia. It is offered as a way to prove genuine biological relations when people are seeking family reunification but lack credible documentary evidence. DNA testing can be an empowering technology for families that have, for whatever reason, lost access to identifying documents. But while it is used by the Department of Home Affairs to corroborate parentage, very occasionally it is also used to verify the legitimacy of a marriage, if the couple are suspected to be siblings.
There is no legislation outlining DNA testing for immigration, meaning that an officer cannot legally demand an applicant take a test even if they suspect fraud. But for those seeking family reunification from countries where documentation is not considered adequate, DNA testing is becoming a de facto obligation. If they refuse to take the test, their application may not be approved. For applicants from wealthy, Western countries, family reunification is almost always granted via legal documents. For applicants from poorer, non-western countries – in particular from African countries – DNA is becoming a nominally optional but functionally necessary hurdle.
After receiving the request for DNA, Tadese and Abebe submitted samples to a laboratory in Melbourne. In November 2012, Tadese received a letter from the lab stating that there was “moderately strong evidence” that he and Abebe were half-siblings. This made no sense. Their families were from entirely different parts of the country and met for the first time at the wedding. “We were shocked,” Tadese tells me. The test results were also sent to immigration authorities who subsequently turned down the visa application on the basis that the marriage was incestuous and therefore void under Australian law.
Tadese has not lived in the same country as Abebe since. Living alone in the suburb of West Footscray, Tadese estimates that he has spent around $50,000 in legal fees fighting the decision. Immigration authorities and the courts have remained adamant that DNA testing is an unbiased way of validating claims of kinship – a technology of reunification. For Tadese it has been just the opposite.
Before meeting ABEBE, Tadese had been married once before. In 2007, when he was still living in Ethiopia, Tadese was introduced to an Ethiopian-australian woman. After marrying they relocated to Melbourne and settled in the south-western suburb of Werribee, a place vastly different from Huruta, a small town in the heart of Ethiopia where Tadese grew up. Before he left Ethiopia he had been working in his father’s retail business, and was hoping to expand it to Australia. Instead, he was working long shifts at a plastics factory. After two years, Tadese’s first marriage broke down. He moved to West Footscray and spent much of his spare time at a nearby Ethiopian Orthodox church. At a service in 2011, a friend suggested that Tadese get in touch with a young woman who was working as a nanny in Cairo. Within a year, he and Genet Abebe were married. “We were ready to change our lives together,” Tadese says. Then they got their test results.
After the laboratory informed them that their relationship was deemed incestuous, Tadese and Abebe consulted their families. There was no historical connection between the Tadeses of Huruta and the Abebes of Addis Ababa. This was confirmed by the priest who had presided over their wedding; he assured the couple that there are strict guidelines in the Ethiopian Orthodox church to prevent incest. Tadese sent a letter to the department stating that he and Abebe were certain that
Tadese sent a letter to the department stating that he and Abebe were certain that there must have been an error during sample collection or testing.
there must have been an error during sample collection or testing. Perhaps, he suggested, there was a distant cousin relationship that had been misconstrued as half-siblingship. He also enquired into the possibility of taking another test.
In response, a senior visa processing officer wrote to the laboratory that performed the original test. A scientist at the lab clarified that the amount of DNA shared between Tadese and Abebe meant that a cousin relationship was possible, but a far lower likelihood than half-siblings. If the same test was undertaken again, they added, the couple would likely get the same result. The delegate from the department refused Abebe’s application on the grounds that she and Tadese were in a “prohibited relationship” under Australian law.
In early January 2014, Tadese travelled to Addis Ababa to spend a few months with Abebe. It was the first time they had seen each other since the wedding. Abebe soon became pregnant. Tadese returned to Melbourne, hired immigration lawyer Godson Nwankwo and lodged an application to what was then called the Migration Review Tribunal, to re-assess the department’s decision. Tadese and Nwankwo collated as much documentation as possible to support his case. Along with affidavits from his and Abebe’s parents, Tadese got a letter from the local government of his home town confirming that the Abebes had never lived in Huruta. “I was confident that we had put together a good case,” Nwankwo tells me.
Tadese’s case wasn’t heard at the tribunal until February 2015, which meant he was unable to be with Abebe when she gave birth to their son, Natnael. At the hearing, the panel suggested that Tadese could resolve the issue quickly by taking another DNA test with a certified laboratory. But the process was expensive and logistically complicated (applicants have to organise and pay for their own tests). Abebe, now living with their infant son in Ethiopia, visited a specimen collection
Tadese’s case wasn’t heard at the tribunal until February 2015, which meant he was unable to be with Abebe when she gave birth to their son.
agency in Ethiopia. According to Tadese, they told her they would not take her sample unless authorities in Australia provided a letter. Tadese’s lawyer sent a letter to the tribunal requesting this authorisation and a time extension. The tribunal seemingly ignored the request, instead affirming the department’s original decision.
With the tribunal hearing over, Tadese considered his options. It seemed to him that the only way out of the predicament was to take another test with the same lab, and then petition the tribunal to review its decision. But how could he be certain that the test wouldn’t come up with the same result? After all, it had been wrong the first time. If he didn’t take the test, though, he knew that it would look like he was actively hiding something. Tadese had, in fact, already arranged a private genetic test for himself, Abebe and Natnael, this time with a different lab. The results confirmed that Natnael was his and Abebe’s biological son. But, again, the results stated that he and his wife were half-siblings. “I didn’t understand what was happening,” he says. “The tests were lying but I didn’t know why. I thought no one would believe me.”
The test that Tadese and Abebe received is what’s known as short tandem repeat (STR) analysis. An STR is a DNA sequence that repeats itself many times in specific locations in the human genome. All of us have these repetitive sequences but the number of times they repeat varies between individuals, functioning like a genetic fingerprint. We inherit these unique sequences from each of our parents, which is why they can be used to determine relatedness. In the original test, the laboratory compared DNA patterns at 21 locations on Tadese and Abebe’s genome. They shared enough repeating patterns for the lab to conclude that they were 66 times more likely to be half-siblings than not. Tadese, who had not encountered DNA testing before, was troubled and confused by the results. Nwankwo, his lawyer, had likewise never encountered genetic testing for siblingship in an immigration case before. “I had never seen anything like it,” he tells me. “Neither had any of my colleagues.”
Nwankwo requested that the tribunal review the decision and, after much delay, the case was heard again in March 2017. Nwankwo elected to include the DNA evidence from Tadese’s private test, but only that which indicated that Natnael was his and Abebe’s biological son. The tactic backfired. The tribunal now knew that Tadese had undergone DNA testing privately but was withholding part of the results. The tribunal held again in favour of the department.
Not knowing where else to turn, Tadese approached his priest who put him in contact with Haileluel Gebre-selassie, a businessman and prominent member of Melbourne’s Ethiopian community, who worked for many years at the Migrant Resource Centre. “When he told me his story I felt that it was almost unbelievable,” Gebre-selassie tells me. He emailed Benjamin Millar, a journalist covering Melbourne’s western suburbs who then published an article in The Age. It was syndicated to a New Zealand news website, stuff.co.nz, which is where Andrew Veale first learnt about Tadese’s predicament.
A senior scientist at Auckland’s Manaaki Whenua Landcare Research facility, Veale had been involved in research calculating genetic relatedness between invasive animal species. Over the years he had occasionally encountered false positives, often caused by inaccurate data from a “reference population” (the population against which two individuals are compared). For example, while working on a project to eradicate stoats on a small island in New Zealand, Veale realised that the population was genetically distinct from other stoat populations. For this reason, if he tested relatedness between any two stoats from the island against a reference population of stoats from other parts of New Zealand it might look as if they were siblings or even parent and child, even if they weren’t. He wondered if a similar error had been made when comparing Tadese and Abebe.
Veale contacted Millar, who then put him in touch with Gebre-selassie and Tadese. Veale asked Tadese to send him the original DNA report issued in 2012. “I read it and immediately there were red flags,” Veale tells me. To begin with, the methodology used to calculate relatedness was from a research paper published in 1996 with only 73 citations. “Given how quickly the field has progressed, this is shockingly out of date,” Veale says. More fundamental, though, was an issue with the reference population data.
Just like with Veale’s stoats, when measuring relatedness between humans, the composition of the reference population data directly impacts the accuracy of the test. This is because certain STR variations are more common in some populations than in others. People of Ethiopian ancestry, for example, may be more likely to share certain DNA patterns that are absent in other populations. If an inaccurate reference population is used to test two individuals of the same ancestry, their relatedness may be overestimated. Or to put it slightly differently, if you are comparing two individuals who share a unique ancestry with a reference population of people from a different ancestry, STR analysis might confuse ancestral similarity for familial relatedness.
Having an accurate reference population is, for this reason, extremely important in STR testing, especially for people of African ancestry. Because our
“I read [the original DNA report] and immediately there were red flags. Given how quickly the field has progressed, this is shockingly out of date.”
species, homo sapiens, first evolved in Africa and lived on the continent in separate groups for many thousands of millennia before migrating elsewhere, significant genetic diversity among those groups emerged over time. Humans who left Africa, in comparison, passed through a population bottleneck. This is called the founder effect and it is still visible in population genetics today. For a geneticist, a random Japanese person is more similar to a random German person than, say, an East African person and a West African person. The possibility for error when not using the right reference population is thus higher when testing kinship in people of African ancestry than it is for others – particularly in Ethiopia, where this vast genetic diversity is most pronounced.
When Veale read Tadese’s original report, he discovered that the laboratory had used a reference population listed as “African” – a very broad and vague category. Veale emailed the company to enquire what they meant by “African” but received no further information. He guessed that the “African” database was likely made up of samples from African-american people, who are mostly of West African ancestry. Such a database, Veale tells me, would be very inaccurate in testing people of Ethiopian ancestry. “Not only is using this database incompetent, it’s also racist,” he says. “It presumes that all Africans are the same when really there is more genetic variation among Africans than anywhere else in the world.”
Veale presumed that the broad reference population likely accounted for why Tadese and Abebe had been found to be half-siblings. Because they are both Ethiopian, and in fact share the same ethnicity within Ethiopia, they would share repeating DNA sequences that are rare or absent in other African populations. “I would wager that any two Ethiopians assessed against this general ‘African’ reference population would appear highly related,” Veale says. “If my undergraduate students
“I would wager that any two Ethiopians assessed against this general ‘African’ reference population would appear highly related.”
made these fundamental errors in assessing genetic relatedness I would fail them.”
To test his hypothesis, Veale arranged for Tadese, Abebe and their son to take another DNA test, this time with the highly popular 23andme service. Like many other modern direct-to-consumer genetic testing services, 23andme does not use STR analysis; it undertakes single nucleotide polymorphism (SNP) analysis to assess relatedness. According to Veale, SNP analysis is a far more powerful and accurate technology, as it can compare DNA at hundreds of thousands of locations on the genome instead of just 21. When the results came back in early 2019, they were unequivocal. Tadese and Abebe shared such little DNA that Veale couldn’t even track a shared common ancestor back four generations, let alone a shared parent. “The lab had made an error,” Veale says. “And Daniel and his family suffered.”
After the tribunal’s decision in 2017, Tadese appealed to the Federal Circuit Court. His case wasn’t heard until May of last year. By then, Tadese had sound scientific evidence showing that he and Abebe were not related. Yet because 23andme was not a laboratory certified by the National Association of Testing Authorities the results were not taken into consideration. Instead, the judge relied on the original 2012 DNA results to inform her decision and again ruled in favour of the department.
For the first time, Tadese felt hopeless. He spent much of his time driving an Uber, saving up to buy airline tickets for his annual visits to Addis Ababa. There, with his son, Tadese’s lingering depression would lift, only to return the moment his plane touched back down in Melbourne. Kafkaesque is an adjective often used offhandedly to describe frustrating encounters with bureaucracy. But there is no word more apt to describe Tadese’s situation. He was stuck in a surreal world in which all of his plans, all sense of control, had fallen to pieces. He found himself pitted against a system that insisted there was some essential and inescapable part of his identity that was to blame for his own predicament. He didn’t want to give up, but he also felt like he didn’t stand a chance.
When DNA testing was first used in immigration department decisions its impact was in stark contrast to what Daniel Tadese experienced. In 1985 when the British geneticist Alec Jeffreys co-published a paper in the scientific journal Nature demonstrating how counting DNA patterns could be used to determine kinship, he received a letter shortly after from a lawyer in London asking if he would assist in a complex immigration case that had been running for two years. A 13-year-old boy had been stopped at Heathrow Airport on his way back to his mother after a trip to visit his father in Ghana. Immigration officials suspected that the boy’s passport was forged. Jeffreys agreed to use his new method to test the boy and his mother to prove parentage. The results plainly pointed to a parent–child relationship. The lawyer presented the data to the Home Office Immigration Tribunal; it immediately dropped the case. “I realised then that we were on to something of real use,” Jeffreys told The Guardian. “We had reached out and touched someone’s life.”
The case was covered extensively in the media, and before long Jeffreys was inundated with calls from families bogged down in immigration disputes. By 1995, his method had been used to test 18,000 immigrants, successfully proving blood relationships for 95 per cent of those who had been initially refused entry to the United Kingdom. But if DNA was being used proactively by those seeking reunification, it didn’t take long for governments to realise that they could also use it as a mechanism of exclusion. There was a growing trend towards strictly enforcing identity requirements for those seeking family reunification in Western countries. While documents could be easily forged, DNA offered a seemingly foolproof way of assessing the validity of a kinship claim.
By 2000, immigration departments in the United States, Canada, the UK, Germany and Finland, among others, were routinely requesting genetic testing from those seeking family reunification. Australia quickly followed: in 2001, the immigration department requested about 200 genetic tests in family immigration cases. Crucially, this biometric validation system was not legislated. Instead, policy guidelines were laid out in the Procedures Advice Manual, a weighty document issued to immigration officials as a companion to the legislation. The guidelines stated that if credible documentation was not provided to prove family relations, DNA testing could be offered as a “last resort” tool in establishing the bona fides of a claimed relationship. The test was not compulsory: if an applicant refused, the guidelines specified that “little weight should be given to this decision”.
Despite these caveats, for those whose documentation had been deemed inadequate, a DNA test was an elective but often necessary hurdle for proving kinship. The Procedures Advice Manual itself made this implication, stating that testing was a “self-selection” procedure as “non-genuine applicants tend not to proceed with DNA testing”.
He assumed the test was a requirement and he trusted the process. “I thought that in this country the systems worked. But then this happened to me.”
In 2003, the Australian Law Reform Commission reviewed the kinship testing policies and raised concerns about “the adequacy of protections afforded to applicants”. Specifically, a question was asked about how “the framing of letters offering testing might lead some applicants to regard testing as a requirement, rather than as an opportunity to provide additional evidence”. But in 2012, when Tadese received his letter from the department requesting DNA, the policy remained the same. He assumed the test was a requirement and he trusted the process. “I thought that in this country the systems worked,” he says. “But then this happened to me, and it shattered that.”
The question of how many other people the system has failed is a hard one to answer. Making DNA tests legally non-compulsory allows the department a measure of distance from the process. Applicants have to take responsibility for the testing themselves and engage the services of a laboratory. The science is outsourced, the methods effectively hidden by proprietary laws. (According to an email from the Department of Home Affairs, DNA testing is “only used in a very small proportion of migration cases”.)
What is known, however, are those who are most frequently targeted for testing. According to Torsten H. Voigt, a professor of sociology at the RWTH Aachen
The judge acknowledged the “credibility” of Tadese and Abebe’s relationship … then dismissed Tadese’s claim and ordered him to pay costs.
University in Germany who has tracked kinship testing for immigration for almost a decade, the use of DNA is creating a prejudicial schism in the family immigration experience. After interviewing several private laboratories that undertake kinship testing in Germany, Voigt estimated that immigration officials use DNA testing to inform decisions in up to 20 per cent of family reunification cases. “Given that most cases of family immigration from Western countries will not require DNA, we can assume that most of these cases involving DNA are from countries with refugees,” Voigt tells me. Carina Ford, managing partner of one of the largest immigration law practices in Australia, says that this dynamic exists here, too. “With applicants from certain African countries, we will often put them on notice that there’s a high chance that DNA testing could be requested,” she says. “I do feel sometimes that DNA unfairly targets certain groups of people.”
If certified laboratories in Australia are using inadequate reference population data, as was the case for Tadese and Abebe, the bitter irony is that those who are most likely to be tested – African applicants – are those who are most likely to suffer from inaccurate results. This might simply be a case of incompetence, one that could be easily rectified by upgrading kinship testing methods to reflect up-to-date science. The fact that it hasn’t, though, mirrors a racialised history of Australian immigration policy, if not specifically in intent then certainly in outcomes.
I tried to contact the laboratory that administered Tadese’s first test, to see whether its database has been upgraded. I received an email in response stating that for privacy reasons it could not discuss the case in question. I followed up with some general questions about the immigration kinship testing process, which the company still conducts, but repeatedly had my requests “escalated to management”. Eventually, I received a call from an unnamed employee at the company asking me to stop contacting staff because it was “becoming too much”. Instead, I contacted several other certified laboratories that perform kinship testing for immigration. When I asked about what reference population data they use to test people of African ancestry, various managers responded that they had a “general” African database, or in one case, a single “global” database. I asked whether this put some people, particularly those from the African continent, at risk of being found as false positives. “This could absolutely happen,” one manager said.
Veale tells me that this could explain why Tadese and Abebe received two false positive results from STR testing in Australia. “[The companies] invested in this in the 1990s and never bothered to update their practice,” he says. “No one challenged them on why they were using outdated technology that would cause problems for people from other countries.”
His assessment was confirmed by two of Australia’s leading geneticists, Stephen Leslie and David Balding, both from the University of Melbourne. When I sent them a copy of Tadese’s original 2012 test, they were shocked that it had been used as the basis of an immigration decision – one that has meant a man and his family have spent nine years unable to live together in the same country. “My biggest question is why in 2012 the immigration department is using an outmoded data type and a test, which appears on the surface to be not possibly applied fairly in immigration cases,” Leslie says. Balding concurs: “They’ve been doing the same tests for 20 years and they are out of date. And if they’re still doing this type of testing in immigration, they’ll be getting a lot wrong.”
Since being dismissed by the tribunal in 2017, Tadese’s case has been heard twice in the Federal Circuit Court. At the first hearing, the judge acknowledged the “credibility” of Tadese and Abebe’s relationship. “On the usual measures,” the judge wrote, “such as their commitment to each other, having a family together, their financial relationship, the nature of their household and the social aspects of their relationship, the applicant and his wife had a genuine marital relationship.” Then, the judge dismissed Tadese’s claim and ordered him to pay costs of $7467. Tadese filed an appeal, which was heard on May 11 this year. At the hearing, no mention was made of Tadese’s test with 23andme. The judge quickly dismissed the appeal, again ordering Tadese to pay costs.
The next day, at a cafe in the Melbourne suburb of Sunshine, I meet with Tadese and Gebre-selassie, who helps him with interpretation. Tadese was not aware that his appeal had been heard the day before. When I break the news, he takes it with resignation. Gebreselassie suggests that he could write a letter and send it to the federal minister, explaining the situation. Another option, I offer, is to take yet another DNA test, this time at a laboratory with an up-to-date database. This could be submitted along with a new visa claim.
Tadese, who until this point had never once betrayed the incredible stress of his predicament outwardly, puts his hands to the bridge of his nose, shakes his head and exhales deeply.
“You see he is a man who has lost his purpose,” Gebre-selassie says. “He is a man unsure about his future.”
“They’ve been doing the same tests for 20 years and they are out of date. And if they’re still doing this type of testing in immigration, they’ll be getting a lot wrong.”