What the DNA Does and Doesn’t Show in the Soering Case

As you might have guessed, I’m working on a project about the Jens Soering case, which explains all the recent posts. I know this might be a bit dull for those of you who haven’t followed the case, but I’d recommend getting into it — it’s quite interesting. This post will be long and fairly wonky, because the argument is complex. For a basic primer on DNA identification, you might want to consult my last post, and the links inside it.

This post will explain why Jens Soering thinks the DNA and blood-group analysis done by his experts proves his innocence — and why this isn’t correct.

There’s a huge amount of information online on the subject of DNA in the Soering case. Here’s a brief, neutral summary.

1. In 1985, DNA testing was not yet available. The bloodstains found all over the Haysom’s house were sampled and categorized only according to blood type.
2. Derek Haysom had Type A blood, Jens Soering had Type O, and Nancy Haysom had Type AB blood.
3. The two victims, Derek and Nancy Haysom, were cremated shortly after their deaths. There are no direct DNA samples from them, but a profile can be inferred from their children, and from consistent DNA results of samples found at the crime scene (see below).
4. Type O blood was found at the crime scene. The prosecution relied on this fact at trial as being consistent with Soering’s presence, which it was. The defense pointed out that 40-45% of the population has Type O blood, which is also true. Once again, the jury was not hoodwinked or deceived, they heard both sides of the argument.
5. In 2009, the State of Virginia conducted a DNA analysis of the blood samples found in the Haysom’s house. Soering did not request this DNA analysis, as is sometimes reported. The analysis was part of a state program to use DNA to detect potential miscarriages of justice.
6. DNA tests do not reveal blood type. The vast majority of the samples used for the 2009 analysis were degraded and unsuitable for testing. A few of the samples generated partial DNA profiles, but with many elements (loci) missing. The person doing the 2009 DNA test found that all the male DNA samples were consistent, and thus assumed that all of the male blood samples came from a “common male contributor”, presumably Derek Haysom.
7. In 2016, Soering’s team reviewed the original blood-typing documents from his trial. After cross-checking the typing results with the DNA results, they discovered that some of the blood samples found at the crime scene were not consistent with Haysom’s blood type.

Soering and his lawyers hired two experts, Drs. McClintock and Schanfield, to cross-reference the blood type and DNA profile evidence. These experts did not conduct any new testing. They simply reviewed existing documents. Therefore, any recording or bookkeeping errors would necessarily affect their conclusions. They argued as follows:

1. Most of the testable male DNA found in the bloodstains was Type A, and consistent with being from one donor, presumably Derek Haysom.
2. However, there was also DNA from a male recorded as having Type O blood (Soering’s type) but whose DNA was inconsistent with Jens Soering.
3. There was also DNA from a male recorded as having Type AB blood, which is a different type from both Derek Haysom and Jens Soering.
4. Therefore, the DNA and blood-typing evidence did not definitively place Soering at the crime scene. This is undisputed.
5. It also suggests there were two unknown male DNA donors who left DNA at the crime scene — one with Type O blood, and one with Type AB.
6. Therefore, the experts concluded with a “reasonable” amount of certainty that there was at least one, and most probably two, unknown males involved in the murders.

Soering’s experts are paid partisan witnesses, which means they have an incentive to emphasize conclusions which fit the defense’s theory. This is legitimate in an adversary system, and there is no evidence the experts are not acting in good faith. Yet independent experts, as I pointed out before, have rejected the “two unknown male DNA donors” theory. All they say is that the DNA and blood group analysis does not place Soering at the crime scene (which nobody disputes), but also does not show anyone other than the Haysoms was at the crime scene.

Why do the independent experts doubt the “two attackers” theory? After all, if there’s Type AB blood from a male there, doesn’t this absolutely require the presence of some stranger, since Derek Haysom had Type A blood, and Nancy Haysom (Type AB) is female?

To understand the answer, we need to dig deeper into the DNA evidence. For that I will rely on an excellent series of blog posts (g) on a German website which I have referenced before. These are anonymous posts by a user named “Sector 7” on an online forum discussing the Soering case. Obviously posts from an anonymous user on a forum need to be handled with care. However, the reasoning and evidence stand and fall on their own. The poster cites his sources, lays out his reasoning in detail, and seems to know what he’s talking about (but of course if you spot an error, please let me know).

These posts are also, curiously, by far the most detailed and informative skeptical discussions of the Soering DNA evidence in any language. Almost every other news source simply regurgitates Soering’s claim that the blood evidence exonerates him, without explaining why. The most important part of Sector 7’s posts is this chart:

This lists the DNA and blood-type test results for all of the testable samples in the Soering case. ‘BG’ stands for blood group. ‘Loci’ is the number of loci which yielded testable DNA. As I explained in a previous post, 13 loci are tested for DNA identification purposes, and 8 matches are required for a legally-admissible DNA match. AMEL records the sex of the DNA contributor, YX for male, XX for female. The other 13 columns list results for the 13 genetic loci (identified by complex abbreviations) used for DNA identification. As you can see, the samples were extremely degraded, and many of the loci reported no results, or only partial results. 17 of the samples came from a male or males, two from a female or females (the sex for one sample, number 15, could not be specified). The last two turquoise entries are almost certainly from Nancy Haysom. At the end, EH contains the full DNA profile of Elizabeth Haysom.

The samples listed in light blue probably all belong to Derek Haysom — as you can see, the allele pairs in light blue match at all of the loci which delivered results, and are all from samples with Type A blood, Derek Haysom’s blood type. It stands to reason that most of the blood at the scene would be Haysom’s, since he fought vigorously for his life, moving around the house while doing so, and suffered the most wounds. The samples in dark blue are from what Soering’s supporters suggest is hypothetical unknown male attacker #1 (let’s call him HUMA1), who has Type O blood (like Soering) but different DNA from Soering. The samples in pink are hypothetical unknown male attacker #2 (HUMA2) who has Type AB blood (Nancy Haysom’s type, which neither Soering nor Derek Haysom have). Based on these profiles, Soering’s experts suggest HUMA1 and HUMA2 left blood at the crime scene, and therefore must have have been involved in the murders. Notable, Both attackers also had DNA consistent with each other at several loci.

Here’s the chart again, with some added labels for clarity:

But look again at the allele pairs at the loci: Even for the dark-blue and pink HUMAs, all the loci which yielded results reported the exact same allele pairs as Derek Haysom’s. That is, if you didn’t know about the blood type evidence, you would assume that the most likely explanation is that all the DNA came from one male, since all the male samples have consistent DNA markers. This is exactly what the 2009 Virginia state DNA examiner assumed — that all the blood came from one male contributor.

Now, if the DNA test had shown that the blood samples came from males who had DNA profiles inconsistent with Soering and Haysom, that would have be strong proof another male attacker left blood at the crime scene. That would have blown the case wide open. In that case, everybody, including me, would endorse the Soering team’s arguments for a new trial. But that’s not what happened. What happened is that some of the samples with consistent DNA had the “wrong” blood types. It all comes down to how important this fact is, and what explains it. Soering’s supporters say this fact is important, and is explained by the likely presence of blood from at least one HUMA at the crime scene.

Is this convincing?

Not really. Skeptics like the independent analysts and Sector7, point out that there could be many other explanations for the “mismatch” between DNA and blood typing evidence. First, the independent experts pointed to the seriously flawed conditions of deposition, collection, handling, and storage of the samples, which explained the fact that 85% of the samples were unusable. Second, there could have been a clerical error in recording the blood types. Another explanation is that the blood samples could have contained different DNA. Blood, saliva, sweat, mucus, skin cells, and hair follicles can all yield testable DNA. By all accounts, there was a ferocious fight leading up to the Haysoms’ deaths: both put up a struggle in close combat, suffered many defensive wounds, and likely inflicted wounds themselves. Further, the assailant used cloths and mops to mop up and wipe many of the bloody surfaces during cleanup, further mixing samples.

Thus, it’s entirely possible that many samples contained blood and other bodily fluids from many of the people involved. And only random chance determined which of these samples would survive for testing. The blood typing test could, therefore, yield one result, but the DNA test a different one. Put another way, the blood could have come primarily from Soering or Nancy Haysom (explaining the Type O and Type AB blood types), but the only testable DNA which survived in the blood stains came from skin cells, hair follicles, sweat, saliva, or blood contributed by Derek Haysom. And in fact this explanation is also completely consistent with the evidence. Under this explanation, the dark-blue bloodstains in the chart above yielded in 1985 an overall type-match of O (Jens Soering’s type), but the only testable DNA traces left in the sample (skin cells, skin oils, blood, bits of tissue, hair follicles, sweat, saliva) in 2009 were contributed by Derek Haysom. That theory also explains the AB match: the pink blood type (as tested in 1985) is AB (Nancy Haysom’s blood), but the only testable DNA traces in it as of 2009 belonged to Derek Haysom.

And that latter theory explains one highly significant fact which jumps off the above chart: All of the male samples show consistent DNA. Not one locus is out of line. Thus, for the exoneration theory to be true, you would have to posit a theory in which a male intruder with Type O blood left DNA in the house (sample #7, 2FE), and this sample happened to overlap with all or some parts of the (inferred) DNA profile of Derek Haysom at 8 separate loci. As I showed in my previous post, compound small probabilites quickly add up to infinitesimal chances. Even a match at only 3 loci rapidly generates probabilities of 1 in 16,000. Matching at 8 loci generates exponentially higher values — in fact, the legal requirements for a DNA match in the USA are a minimum of 8 matches, generating a probability of at least 1 in 10 million.

It’s difficult to calculate the exact probability of this coincidence (since we’re comparing incomplete profiles), but it’s extremely small. Sector 7, using USA allele-frequency databases, estimates it, as a rough ballpark figure, as 1 in 68 billion. And that doesn’t even take into account the results for the other 3 samples from the two hypothetical unknown male accomplices, which also happen to be 100% consistent with each other, and with the DNA profile of Derek Haysom. Now, because the profiles are incomplete, it’s hard to generate a definitive probability analysis — but it’s accurate to say the chances of both of these unknown males having DNA this similar (and consistent with Haysom’s) — is, even under the most conservative set of assumptions imaginable —  in the 1-in-millions range.

So, to sum up, there are two possibilities here.

Scenario Number One: There were two unknown male intruders, whom nobody has ever been able to locate or identify, who both left DNA at the crime scene. And both of these men happened, by pure happenstance, to have DNA which is completely consistent not only with each other’s DNA, but also with DNA from Derek Haysom at up to 8 separate loci. The chances of this happening are, at the very absolute minimum, 1 in several million, and in actuality probably much lower even than that.

Scenario Number Two: Of the small minority of male DNA samples which survived for testing results, all of these came from Derek Haysom — whether from skin cells, sweat, saliva, mucus, or blood. However, they were found in mixed-up blood samples which contained mostly Type O and Type AB blood, which led to them being classified with the “wrong” blood type (i.e., not Derek Haysom’s Type A). It’s also possible there was some error in testing or recording the blood types.

Which of these scenarios is true? We’ll never know with absolute certainty, unless some new technology develops which can finally resolve this issue. But it’s important to note that Scenario Number One — the “two unknown assailants with consistent DNA” scenario — depends on an extremely improbable 1-in-millions coincidence. So even if you think Scenario Two (mixed samples/human error) is improbable, it is almost certainly much more probable than Scenario One. As Sherlock Holmes once said, once you eliminate theories which are impossible, then any remaining theory which explains the facts, however improbable, must be the correct one. This is a fallacy in certain circumstances, but not here, where there is a limited set of conceivable explanations.

And of course, in line with my last post, the DNA evidence must be considered in the context of all other evidence. A robust finding of male DNA which wasn’t Soering’s or Haysom’s would be a game-changer. That would have got everyone’s attention, and that finding would almost certainly have made Soering a free man by now. But the DNA evidence showed nothing which is inconsistent with Soering’s guilt. And therefore, it doesn’t undermine the mountain of other circumstantial evidence which proved, beyond a reasonable doubt, that Jens Soering, acting alone, killed the Haysoms.