Peripheral blood leukocyte samples were collected from 151 common bottlenose dolphins (Tursiops truncatus) in the course of capture/release health evaluation studies at four different sites: Charleston Harbor, SC; Indian River Lagoon, FL; Sarasota Bay, FL; and St. Joseph Bay, FL. RNA was extracted and hybridized to a first-generation dolphin microarray. We tested the hypothesis that individual dolphins could be assigned to their home regions (by machine learning methods) using only their transcriptomic signatures as classifiers. The machine learning approaches used in this study were artificial neural networks (ANNs) which were able to identify gene expression differences in males and females in some geographical locations. As the sex ratios sampled in each location were not the same and could influence the classification of individuals to locations, males and females at each location were considered separately. ANNs were able to correctly classify dolphins according to their site of sampling with a high degree of confidence. The basis for this result may lie in genetic differences between populations, or in environmental factors (including for example diet, infection, contaminant load, or exposure to biotoxins) or in combinations of these factors. These results suggest that a combination of microarrays and machine learning analytical approaches will be a powerful approach to understanding the interaction of dolphins with the marine environment.
Transcriptome Profiles: Diagnostic Signature of Dolphin Populations
MANCIA, Annalaura;
2010
Abstract
Peripheral blood leukocyte samples were collected from 151 common bottlenose dolphins (Tursiops truncatus) in the course of capture/release health evaluation studies at four different sites: Charleston Harbor, SC; Indian River Lagoon, FL; Sarasota Bay, FL; and St. Joseph Bay, FL. RNA was extracted and hybridized to a first-generation dolphin microarray. We tested the hypothesis that individual dolphins could be assigned to their home regions (by machine learning methods) using only their transcriptomic signatures as classifiers. The machine learning approaches used in this study were artificial neural networks (ANNs) which were able to identify gene expression differences in males and females in some geographical locations. As the sex ratios sampled in each location were not the same and could influence the classification of individuals to locations, males and females at each location were considered separately. ANNs were able to correctly classify dolphins according to their site of sampling with a high degree of confidence. The basis for this result may lie in genetic differences between populations, or in environmental factors (including for example diet, infection, contaminant load, or exposure to biotoxins) or in combinations of these factors. These results suggest that a combination of microarrays and machine learning analytical approaches will be a powerful approach to understanding the interaction of dolphins with the marine environment.I documenti in SFERA sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
