In This Issue

Biodiversity loss in grasslands
Native-dominated (foreground) and exotic-dominated (background) grasslands. Image courtesy of Catherine E. Koehler (University of California Donald and Sylvia McLaughlin Natural Reserve).
The consequences of global warming for species diversity are unclear, given the lack of consistent global trends toward either increased or decreased diversity. Susan Harrison et al. (pp. 8672–8677) monitored plant species at 80 individual sites in a California grassland over a 15-year period from 2000 to 2014. In this and other semiarid regions, forecasts predict that the climate will become increasingly dry over the coming century. The authors found that the number of species declined both within individual sites and among all sites studied. Native annual wildflowers constituted the largest share of the decline, and within this group there was a decline in the mean specific leaf area (SLA). High SLA is associated with low drought tolerance, and the decline in SLA suggests a disproportionate loss of drought-intolerant species. The authors also observed a substantial decline in precipitation during midwinter that was a significant predictor of diversity loss. Other possible causes of diversity loss, such as livestock grazing, fire, nitrogen deposition, and increasing nonnative species appeared to have no significant effect on the number of species. The results indicate the likelihood of diversity loss in semiarid regions as the regions become increasingly dry due to climate change, according to the authors. — B.D.
Analyzing fingerprint recognition accuracy
The unique and persistent pattern of ridges on human fingers has been considered admissible forensic evidence in the United States since a milestone court ruling in 1923. However, belief in the persistence of human finger ridge patterns is largely based on evidence from case studies. Because fingerprint recognition can be confounded by skin changes, technical artifacts, and coincidental matches, Soweon Yoon and Anil Jain (pp. 8555–8560) examined tenprint records, which are cards of fingerprints collected in controlled settings, of 15,597 repeat offenders in a Michigan police database, with five or more records spanning at least 5 years for each person. Using two commercial fingerprint matching kits, the authors analyzed how match scores varied for genuine and impostor pairs with varying time intervals, fingerprint image quality, and age, gender, and race of the individuals. Statistical modeling revealed that although genuine match scores declined with increasing time interval between fingerprinting, fingerprint recognition accuracy remained stable for at least 12 years. Genuine match scores declined with increasing age of the individuals and decreasing image quality, but gender and race had little impact on the scores. Impostor match scores varied little over time and with the individuals’ ages. The findings cannot be directly applied to fingerprints left unintentionally, but they underscore factors influencing fingerprint recognition, according to the authors. — P.N.
Ancestry of Brazilians
Ancestry of EPIGEN individuals, grouped by Brazilian region.
The ancestry of the population of Brazil includes various geographical sources, presenting an opportunity to study the genetic characteristics of population admixture. Fernanda Kehdy et al. (pp. 8696–8701) analyzed data from the EPIGEN Brazil Initiative, which included 6,487 individuals classified according to location within Brazil. The authors found a pervasive pattern of assortative mating throughout Brazilian history in which Brazilians chose others of similar ancestry. Tracing ancestry of Brazilians from the southeast and south regions with predominantly European ancestry suggested a wider source area in Europe and the Middle East than previously supposed. Further, the authors found more recent European migration to the southeast and south regions than to the northeast region, and found that the low Native American ancestry observed in admixed Brazilians was introduced soon after European colonization. Analysis of Brazilians with African ancestry suggested two source areas within Africa, one from non-Bantu populations from West Africa, associated with northeastern Brazil, and one from Bantu groups from East Africa, associated with southeastern and southern Brazil. The results, which include whole genome analysis of 30 people, suggest that the distribution of deleterious genotypes in Brazilian populations likely depends to a greater extent on the amount of continental admixture than on local history factors such as consanguinity or family structure, according to the authors. — P.G.
Human olfactory fingerprints
Human olfactory fingerprints. Image courtesy of Ofer Perl (Weizmann Institute of Science, Rehovot, Israel).
Around 30% of the human DNA that encodes olfactory receptors is unique to each individual. To derive a fingerprint of olfactory perception, Lavi Secundo et al. (pp. 8750–8755) recruited 89 people around 26 years of age and analyzed the participants’ perceived pairwise similarity ratings for 28 odors using 54 descriptors such as “very masculine” and “very lemony.” Though the average description of an odor did not vary among participants, the authors discerned a fine perceptual fingerprint for each individual that was independent of the particular descriptors but varied with the odors used to develop the fingerprint. Reliable olfactory fingerprints could be obtained in less than 10 minutes using only 7 odors and 11 descriptors. When tested 10–30 days after initial fingerprinting, fingerprints of any individual were more similar to the individual’s own fingerprint than to those of another individual. In another group of 130 people around 30 years of age, pairs of individuals who were well-matched in blood proteins called HLA, implicated in immune response and organ rejection, were better matched in olfactory fingerprints than pairs with a poor HLA match, suggesting that olfactory fingerprints might provide clues to immune compatibility. According to the authors, olfactory fingerprints might potentially obviate the need for an estimated 30% of HLA matching tests in population-wide screening. — P.N.
Role of semantic knowledge in reading aloud
The integration of cognitive models of reading with neural data is currently limited. Paul Hoffman et al. (pp. E3719–E3728) used a distortion-corrected functional MRI protocol to examine activation in the brains of 27 participants as the participants read individual words aloud. In addition to the expected activation of brain regions involved in processing written words, a function called orthography, and in processing speech sounds, a function called phonology, the authors observed activation in the lateral anterior temporal lobe (ATL), a brain region associated with processing word meaning, a function called semantics. Greater ATL activation was observed when participants read words with irregular spellings, such as “island,” relative to reading regularly spelled words. Modeling of the interactions between orthographic, phonological, and semantic regions of the brain showed increased involvement of semantic processing when reading irregularly spelled words. Further, 24 participants were also tested on their semantic reliance, or the extent to which an individual depends on semantic knowledge when reading aloud. Differences in lateral ATL activation between individuals correlated positively with semantic reliance. The observations are consistent with the Triangle Model of reading, which proposes that reading aloud involves two neural pathways: a direct visual-to-phonological pathway for words whose pronunciations follow typical patterns, and an indirect pathway mediated by semantic knowledge for pronouncing irregular words. — B.D.
Antibodies help prevent and treat MERS in mouse model
SEM of MERS virus particles on infected cell. Image courtesy of National Institute of Allergy and Infectious Diseases
Since September 2012, the Middle East Respiratory Syndrome coronavirus (MERS-CoV) has sickened more than 1,100 people, with a death toll higher than 400. Development of vaccines or drugs against the virus has been hampered by the lack of physiologically relevant animal models of infection. Using a genetic engineering tool called VelociGene, Kristen Pascal et al. (pp. 8738–8743) replaced the mouse gene encoding a receptor targeted by the virus with its human counterpart and created a mouse model of MERS-CoV infection. Next, the authors used a method called VelocImmune to generate human-compatible monoclonal antibodies against a viral envelope protein crucial to infection. Two of the antibodies—REGN3051 and REGN3048—tightly bound to evolutionarily conserved regions of MERS-CoV, blocked viral entry into susceptible cells, and neutralized infection by a broad range of viral strains more potently than did antibodies generated using traditional methods. More importantly, when administered 1 day before viral infection, both antibodies greatly reduced lung pathology in the mouse models, and when REGN3051 was administered 1 day after viral infection, the antibody blocked viral replication and reduced lung pathology to a greater extent than did mock antibodies. According to the authors, the mouse model and human antibodies might represent promising tools in efforts to prevent and treat MERS. — P.N.