Systematics of Snailfishes (Family Liparidae): Description of Two New Species from the Aleutian Islands and Investigation of the Utility of Exon Capture Techniques

Abstract

The snailfishes of the family Liparidae are a speciose and geographically widespread family of fishes residing in cold and temperate waters worldwide in both the shallow and deep ocean. Characterized by soft, scaleless, gelatinous, frequently slimy bodies, snailfishes are often difficult to work with both morphologically, as specimens are rarely able to be collected in pristine condition, as well as genetically, as there are known difficulties with extracting sufficient high quality DNA from snailfish tissues. Due to these difficulties, as well as generally conserved morphology with differences among species often based on reductive characters, phylogenetic relationships within the family are poorly understood at multiple levels. One major challenge is that there are many undescribed species, and some species complexes require taxonomic revisions. This study addresses taxonomic revision of one group of snailfishes through the descriptions of two new snailfish species from the Aleutian Islands. Both are similar to Careproctus candidus, described originally from four specimens, which is redescribed here on the basis of 67 specimens in addition to the type specimens. All are allocated to the subgenus Temnocora. At the genus level relationships are also poorly understood, as the most recent family-wide phylogenetic analyses recover all three of the most speciose genera as paraphyletic. Most phylogenetic hypotheses for the Liparidae have been inferred based on only morphological characters, only mitochondrial gene sequences, or a combination of the two. Only one phylogeny has been inferred using genome-wide sequences. This study attempts to add to the understanding of the Liparidae phylogeny by inferring a second phylogeny based on sequences taken from across the genome using targeted exon capture. The generated exon capture dataset had widespread cross-contamination. Thus, multiple methods for detecting and attempting to remove that contamination were explored. Methods included comparing topology of well supported published trees to trees inferred based on this dataset, comparing p-distances based on a priori hypotheses coupled with iterative removal of the most contaminated specimens, and using locus heterozygosity to detect contamination. All methods confirmed the presence of widespread contamination in the dataset; however, it was so widespread that it could not be specifically removed. Finally, suggestions are made for how to avoid and or detect and remove contamination in future exon-capture studies.