Domain adaptation addresses the problem where data instances of a source domain have different distributions from that of a target domain, which occurs frequently in many real life scenarios. This work focuses on unsupervised domain adaptation, where labeled data are only available in the source domain. We propose to interpolate subspaces through dictionary learning to link the source and target domains. These subspaces are able to capture the intrinsic domain shift and form a shared feature representation for cross domain recognition. Further, we introduce a quantitative measure to characterize the shift between two domains, which enables us to select the optimal domain to adapt to the given multiple source domains. We present experiments on face recognition across pose, illumination and blur variations, cross dataset object recognition, and report improved performance over the state of the art.

ObjectiveTo determine whether the GGC repeats in the NOTCH2NLC gene contribute to amyotrophic lateral sclerosis (ALS).MethodsIn this study, 545 ALS patients and 1,305 healthy controls from mainland China were recruited. Several pathogenic mutations in known ALS-causative genes (including C9ORF72 and ATXN2) and polynucleotide repeat expansions in NOP56 and AR genes were excluded. Repeat-primed PCR (RP-PCR) and GC-rich PCR were performed to determine the GGC repeat size in NOTCH2NLC. Systematic and targeted clinical evaluations and investigations, including skin biopsy and dynamic electrophysiologic studies, were conducted in the genetically affected patients.ResultsGGC repeat expansion was observed in 4 patients (numbers of repeats: 44, 54, 96, and 143), accounting for approximately 0.73% (4/545) of all ALS patients. A comparison with 1,305 healthy controls revealed that GGC repeat expansion in NOTCH2NLC was associated with ALS (Fisher's exact test, 4/545 vs 0/1,305, p = 0.007). Compared to patients with the neuronal intranuclear inclusion disease (NIID) muscle-weakness-dominant subtype, patients with ALS phenotype carrying the abnormal repeat expansion tended to have a severe phenotype and rapid deterioration.ConclusionOur results suggest that ALS is a specific phenotype of NIID or that GGC expansion in NOTCH2NLC is a factor that modifies ALS. These findings may help clarify the pathogenic mechanism of ALS and may expand the known clinical spectrum of NIID.

No abstract

Autosomal dominant polycystic kidney disease (ADPKD) is a common cause of renal failure that is due to mutations in two genes, PKD1 and PKD2. Vascular complications, including aneurysms, are a well recognized feature of ADPKD, and a subgroup of families exhibits traits reminiscent of Marfan syndrome (MFS). MFS is caused by mutations in fibrillin-1 (FBN1), which encodes an extracellular matrix protein with homology to latent TGF-b binding proteins. It was recently demonstrated that fibrillin-1 deficiency is associated with upregulation of TGF-b signaling. We investigated the overlap between ADPKD and MFS by breeding mice with targeted mutations in Pkd1 and Fbn1. Double heterozygotes displayed an exacerbation of the typical Fbn1 heterozygous aortic phenotype. We show that the basis of this genetic interaction results from further upregulation of TGF-b signaling caused by Pkd1 haploinsufficiency. In addition, we demonstrate that loss of PKD1 alone is sufficient to induce a heightened responsiveness to TGF-b. Our data link the interaction of two important diseases to a fundamental signaling pathway.