Angie Ambers, Ph.D., Assistant Director, Henry C. Lee Institute of Forensic Science

Case history

Born in 1643, the French explorer Robert Cavalier (Sieur de La Salle) is best known for leading an expedition from the Great Lakes region of the United States and Canada along the Mississippi River, discovering the mouth of the river as it flowed into the Gulf of Mexico. In 1682, he claimed the Mississippi river basin for France and named it ‘Louisiana’ in honor of King Louis XIV. After returning to France in 1683, La Salle was commissioned by the king to sail back to the New World (North American continent), establish a French colony at the mouth of the Mississippi River, and to do reconnaissance on a possible later French invasion of Spain’s silver mines in Nueva Vizcaya (modern-day northern Mexico). In the summer of 1684, La Salle departed France with a large expedition of four hundred sailors and colonists aboard a fleet of four ships (Le Joly, L’Aimable, Le Saint-Francois, and La Belle).

In September 1684, the Le Saint-Francois was seized by privateers and only three ships made it to the French Caribbean port of Petit-Goave (present-day Haiti), a planned stop along the journey to the New World. After a few months of respite at Petit-Goave, La Salle’s remaining expedition sailed towards the Gulf of Mexico. In 1685, the fleet arrived at the western Louisiana coast and proceeded to navigate along the Texas coastline using rough geographical maps previously constructed by La Salle and his mapmaker. The expedition mistakenly landed at Matagorda Bay (near the city of Victoria, Texas), approximately four hundred miles southwest of their intended destination (the mouth of the Mississippi River). At this point in the journey, Le Joly (a military escort ship) returned to France. The remaining two ships, L’Aimable and La Belle, entered Matagorda Bay because the area offered protection from rough seas, pirates, and Spanish soldiers. L’Aimable (a much larger ship than La Belle) ran aground navigating the pass. La Belle successfully entered and anchored in the bay; La Salle and some of his men then explored the surrounding terrain, searching on foot for the mouth of the Mississippi River.

Rene-Robert Cavelier, Sieur de la Salle. Credit: Britannica

Provisions from La Belle and Le Joly (removed from the latter before it set sail back to France), as well as some cargo salvaged from the wreck of L’Aimable, were used to forge a temporary encampment along the coastline at the entrance to Matagorda Bay. Throughout the year 1685, La Salle continued searching for the mouth of the Mississippi River and began building a French settlement thirty miles (~48 kilometers) inland, a colony known as Fort St. Louis. In early 1686, La Belle was caught in a storm and sank to the bottom of the bay, along with the many provisions for the French colony. Within a year, the colonists were struggling due to shortage of supplies and food. In January 1687, La Salle left on foot with sixteen men to seek help. During the journey, La Salle was murdered by his own men, and the local Karankawa Indians massacred the remaining colonists at Fort St. Louis (except for some of the children, who were taken to be raised with the tribe). La Salle’s expedition chronicler, Henri Joutel, made it to Canada and sailed back to France; his journal details the events of the ill-fated expedition.

Spain learned of the French attempt to establish a colony in the part of the New World they controlled and authorized eleven expeditions to find La Salle’s settlement. In 1689, Spanish explorer Alonso de Leon arrived at the Texas coastline and discovered the remains of La Salle’s French colony; another expedition found the partially submerged remains of La Belle and sketched the approximate location of the shipwreck. For three centuries, the ship remained at the bottom of the Matagorda Bay. In the late 1970s, using the seventeenth-century Spanish map, Henri Joutel’s expedition journal, and a magnetometer (a remote-detection instrument similar to a highly sensitive metal detector), the Texas Historical Commission (THC) attempted to locate the shipwreck. Early magnetometer searches involved dragging a magnetometer sensor behind a search vessel and recording readings relayed from the sensor to a receiver in the cabin of the boat. The premise of using a magnetometer in the search relies on the fact that most shipwrecks contain ferrous metal (e.g., in the form of bolts or other fittings that hold wooden timbers together). Intensive surveys of the waters in the Matagorda Bay were conducted, but ended without success.

Over time, technological advances in marine archaeology improved the investigators’ chances of locating the historical shipwreck. The THC resumed its search for La Belle using an improved magnetometer coupled to a Global Positioning System (GPS), which used satellites to accurately track magnetic readings and their respective positions in the water. In 1995, after twenty years of searching, THC divers investigating the source of a magnetic signal near the Matagorda Peninsula discovered a bronze cannon bearing the insignia of The Count of Vermandois (admiral of France from 1669-1683 and illegitimate son of King Louis XIV), confirming that La Belle had been found.

With support from the Texas legislature, $1.75 million was appropriated for excavation of the La Belle shipwreck. Numerous Texas philanthropists and private foundations subsequently donated additional funding to augment the government’s allocation. In 1996, a double-walled metal cofferdam was constructed around the shipwreck and water was pumped out of the site, allowing for a dry excavation of La Belle and its contents. This was the first de-watered cofferdam excavation conducted in the Western Hemisphere, and more than 1.8 million artifacts were recovered. The remaining portions of the hull of La Belle has since been restored and is on display at the Bullock Texas State History Museum in Austin, Texas, along with numerous artifacts recovered from the shipwreck site.

Among the artifacts discovered by the archaeology team were human skeletal remains. “Individual One” was a partial skeleton (right fibula, one metatarsal, two foot phalanges, two hand phalanges) recovered from alongside cargo in the rear portion of the ship. A complete skeleton of an adult male (“Individual Two”) was found on top of coiled anchor rope in the bow of the ship. Early attempts (late 1990s) at DNA recovery were unsuccessful. Numerous advances in DNA technology have occurred since the discovery of La Belle. In late 2015, DNA testing commenced on the skeletal remains. To maximize the amount of genetic information that could be recovered from the remains, a broad range of DNA markers in the human genome were targeted for analysis, including autosomal short tandem repeats (STRs), Y-chromosome STRs (Y-STRs), X-chromosome STRs (X-STRs), ancestry-informative single nucleotide polymorphisms (aiSNPs), phenotypic-informative SNPs (piSNPs), identity-informative SNPs (iiSNPs), and mitochondrial DNA (mtDNA). One female analyst conducted all testing on the skeletal remains of both “Individual One” and “Individual Two,” including surface cleaning and decontamination, bone pulverization, DNA extraction, DNA quantification, PCR amplification, Y-STR genotyping (CE), massively parallel sequencing (MPS) (ForenSeq™ panel), and whole mitochondrial genome sequencing.

“Individual One” – Discovered alongside cargo in the rear portion of La Belle

Traditional Y-chromosome (Y-STR) typing via capillary electrophoresis (CE)

The right fibula recovered from the rear cargo section of La Belle was submitted for testing. Diagenesis (decomposition) of bone microstructure and degradation of DNA within the bone matrix does not occur uniformly; therefore, given the advanced age of these remains, multiple sections along the diaphysis of the fibula were sampled for intact/amplifiable DNA. Extractions were performed on eleven different cuttings. DNA was not detected in extraction reagent blanks or amplification negative controls. Positive controls yielded the correct haplotype for all amplification reactions.

In January 2016, during the first phase of testing, DNA extractions were performed on four cuttings from the diaphysis of the fibula. Initial Y-STR typing was performed using the Yfiler™ PCR amplification kit, a multiplex that simultaneously amplifies 17 different Y-STR markers. This kit contains the core Y-STR loci advocated by the Scientific Working Group on DNA Analysis Methods (SWGDAM) and the European minimal haplotype. Total DNA recovery (ng/µL) from each cutting and total input DNA (ng) for the Yfiler™ amplification reactions were 0.002ng/µL-0.279ng/µL and 0.024ng-1.000ng, respectively. Partial-to-full 17-locus Y-STR profiles were obtained from each fibula section, and allele calls between all samples were concordant. A composite (consensus) Y-STR haplotype was generated and was compared to 1) reference samples from the archaeological team who excavated and handled the remains, and 2) the positive control (male DNA 007) included in the amplification kit. Each archaeological team member’s haplotype differed from the consensus fibula profile at 8, 13, and 11 of the 17 Y-STR loci, respectively. The haplotype of the male positive control DNA was discordant at 12 loci from the profile obtained from the fibula. These non-matches exclude the archaeological team and the positive control DNA as being the source of the profile.

Table 1: Y-STR typing results for four cuttings (001-004) from the right fibula of an adult male (“Individual One”) recovered from the La Belle shipwreck, using the Yfiler™ PCR amplification kit. Allele calls between all samples were concordant and a composite/consensus Y-STR haplotype is reported. The Y-STR haplotype obtained from the fibula was compared to 1) reference samples from the archaeological team who excavated and handled the remains, and 2) the positive control (male DNA 007) included in the amplification kit. Shaded alleles represent non-matches to the haplotype of the unidentified remains, excluding the archaeological team and positive control DNA as being the source of the profile.

In a subsequent phase of testing, seven additional cuttings (005-011) were taken from the fibula diaphysis and extracted for DNA. For these samples, DNA quantification was performed with Quantifiler™ Trio, which quantifies three different loci and provides information about the level of degradation in each sample. Using the Y-target quantity, Y-STR typing was performed with Yfiler™ Plus, an enhanced multiplex that is purported to increase the discriminatory power of Y-haplotype data, and which has been shown to increase typing capability of Y-STRs in challenging bone samples. The Yfiler™ Plus kit includes the same 17 Y-STR markers from the original Yfiler™ kit plus 10 additional highly polymorphic markers (DYS449, DYS460, DYS481, DYS518, DYS533, DYS570, DYS576, DYS627, and DYF387S1a/b) [20,42]. Total input DNA for Yfiler™ Plus amplifications and degradation indices (DIs) ranged from 0.012ng-0.212ng and 2.097-6.986, respectively. Y-STR typing results are shown in Table 2.

Table 2. Y-STR typing results for seven cuttings (005-011) from the right fibula of an adult male (“Individual One”) recovered from the La Belle shipwreck, using the Yfiler™ Plus PCR amplification kit. Each cutting yielded one powder fraction (labeled 005.001 through 011.001, E1 = elution #1). Shading denotes loci that are shared (common) between the Yfiler™ and Yfiler™ Plus kits. Allele calls between all samples were concordant and a partial composite/consensus Y-STR haplotype is reported.

Y-chromosome (Y-STR) typing via massively parallel sequencing (MPS)

Fifteen of the 24 Y-STR markers amplified with the ForenSeq™ kit overlap with the Yfiler™ kit; 19-out-of-24 Y-STR markers in the ForenSeq™ panel overlap with the Yfiler™ Plus kit. Among the common loci between the MPS and CE kits used, Y-STR alleles recovered from all bone sections were concordant. Using a ≥10x read depth threshold, Y-STR typing results were obtained for 18 of the 24 markers assayed with MPS; read depth ranged from 11x to 2361x [340x ± 496x (Avg ± SD)].

Y-chromosome Haplogroup assignment

Y-STR data for “Individual One” were imported into the Y-STR Haplotype Reference Database (YHRD, www.yhrd.org) for biogeographic ancestry assessment using the database’s ancestry feature and metapopulation tool. No matches were found between the Y-STR haplotype obtained from the fibula and the 209,111 Yfiler™ haplotypes contained in the YHRD database. An additional search was performed using the expanded Yfiler™ Plus profile; no matches were found when compared to the 45,892 Yfiler™ Plus haplotypes in the database.

Using Haplogroup Predictor (http://www.hprg.com) and the “equal priors” assumption, both the composite 17-locus Yfiler™ and 27-locus Yfiler™ Plus haplotypes were determined to be consistent with Haplogroup Q. Haplogroup Predictor predicts Y-chromosome haplogroups from Y-STR data based on "fitness score" calculations and Bayesian probability assessment. For the Y-STR profile obtained from the fibula, fitness scores and Bayesian probabilities for Haplogroup Q were 62 (100%) and 63 (100%) using Yfiler™ and Yfiler™ Plus haplotype data, respectively.

Prediction of Haplogroup Q for “Individual One” was an interesting finding. Given that La Belle sailed to the New World from France, the predominant assumption was that all skeletal remains recovered from the shipwreck would be European in origin. Haplogroup Q is the primary paternal lineage of Native Americans; approximately 90% of pre-Columbian Native Americans belong to this haplogroup. In Europe, Haplogroup Q is distributed in low frequencies. There are small populations of males in Europe who possess Haplogroup Q, including in southern Sweden (5%), among Ashkenazi Jews (5%), and throughout various isolated pockets in central and Eastern Europe (e.g., the Rhône-Alpes region of France, Sicily, Poland, Ukraine, Croatia, Serbia).

Ancestry-informative SNPs (aiSNPs)

A panel of 56 ancestry-informative SNPs (aiSNPs) were tested to further investigate the ancestry of “Individual One.” Results were obtained for all 56 aiSNPs, and read depth ranged from 10x to 4313x [321x ± 546x (Avg ± SD)]. Table 3 includes the top ten most likely populations of ancestral origin for this individual, as determined via comparison to the FROG-kb database, which currently contains aiSNP data for more than 164 reference populations. These data and the aforementioned Y-STR data support Native American ancestry for “Individual One.”

Table 3. Population likelihoods for “Individual One” (adult male, La Belle shipwreck) based on aiSNP data obtained from the right fibula and comparison to 164 reference populations in the FROG-kb database.

During archaeological excavation of the La Belle shipwreck, it was presumed that any recovered human remains were European (i.e., most likely French) in origin. However, genetic testing indicated that the fibula recovered from rear portion of the ship likely originated from an individual of Native American ancestry. This led to discussions regarding how an adult Native American male ended up in a sunken French ship. According to the official ship log and diary accounts of Henri Joutel, La Salle’s expedition chronicler (who survived and eventually returned to France), a Native American individual indeed was aboard La Belle when it departed Europe. This person, known as Nika, was a Shawnee Indian that had been given to La Salle on an earlier expedition in North America. However, the Native American remains recovered from La Belle cannot belong to Nika, as he was murdered in 1687 along with La Salle (according to historical records and diary accounts). More likely, the remains are of a local Karankawa Indian who was scavenging the shipwreck, became trapped, and perished. The latter seems plausible due to the shallowness of the Matagorda Bay (4 meters), as well as the fact that the remains were found in the cargo portion of the ship. It is known that King Louis XIV funded La Salle’s expedition with the partial intention of having a French colony established on the Gulf of Mexico; as such, La Belle was loaded with provisions for the journey as well as with supplies to support colonization.

Phenotypic-informative SNPs (piSNPs)

Current MPS technology enables investigators to predict certain phenotypic traits of unidentified human skeletal remains. A panel of 22 phenotype-informative SNPs (piSNPs) were tested. Results were obtained for 22/22 piSNPs, and read depth ranged from 10x to 666x [179x ± 167x (Avg ± SD)]. Using these data and the HIrisPlex hair/eye color prediction tool (http://hirisplex.erasmusmc.nl), “Individual One” most likely had black hair and brown eyes, with probabilities of 0.663 and 0.997, respectively. This phenotype prediction is consistent with the bioancestry data.

Additional DNA markers assayed with the ForenSeq™ MPS panel

The ForenSeq™ panel also assays autosomal STRs, X-STRs, and identity-informative SNPs (iiSNPs). Results were obtained for 22/27 autosomal STRs (plus amelogenin), 4/7 X-STRs, and 94/94 iiSNPs. Read depth for autosomal STRs, X-STRs, and iiSNPs ranged from 10x to 6063x [430x ± 679x (Avg ± SD)], 10x to 463x [114x ± 125x (Avg ± SD)], and 10x to 3589x [256x ± 377x (Avg ± SD)], respectively.

Mitochondrial DNA (mtDNA) analysis

Sequence results for fifteen DNA samples extracted from a fibula of “Individual One” displayed an advanced level of mtDNA degradation, in the form of single base damage and a reduced number of reads. Several positions revealed mixtures of C/T and G/A, suggesting the presence of oxidative and/or hydrolytic damage of the DNA. This phenomenon, however, is frequently observed in ancient and damaged templates. The presence of such damage supports the authenticity of the extracted DNA as originating from these historical remains, as opposed to being the result of amplification of modern DNA contamination. Samples with single read depth or no representation in several regions of the mitogenome were rejected from further analysis. The consensus mitotype was assembled from sequence data of two samples with the highest read depth, and the whole mtDNA genome (1-16569) was considered for haplogroup assessment. Read depth for fibula cutting 001.001 was 977x to 38378x [8158x ± 4738x (Avg ± SD)]. For a second sample (fibula cutting 003.002), read depth was 40x to 17509x [1443x ± 1955x (Avg ± SD)]. Results assign this set of remains to Haplogroup C1b, which is considered to be a pan-American mitochondrial haplogroup and indicates Native American ancestry in the maternal line of “Individual One.”