Background Taste buds contain ~60 elongate cells and several basal cells. Elongate cells comprise three functional taste cell types: I - glial cells, II - bitter/sweet/umami receptor cells, and III - sour detectors. Although taste cells are continuously renewed, lineage relationships among cell types are ill-defined. Basal cells have been proposed as taste bud stem cells, a subset of which express Sonic hedgehog (Shh). However, Shh+ basal cells turnover rapidly suggesting that Shh+ cells are precursors of some or all taste cell types. Results To fate map Shh-expressing cells, mice carrying ShhCreERT2 and a high (CAG-CAT-EGFP) or low (R26RLacZ) efficiency reporter allele were given tamoxifen to activate Cre in Shh+ cells. Using R26RLacZ, lineage-labeled cells occur singly within buds, supporting a post-mitotic state for Shh+ cells. Using either reporter, we show that Shh+ cells differentiate into all three taste cell types, in proportions reflecting cell type ratios in taste buds (I > II > III). Conclusions Shh+ cells are not stem cells, but are post-mitotic, immediate precursors of taste cells. Shh+ cells differentiate into each of the three taste cell types, and the choice of a specific taste cell fate is regulated to maintain the proper ratio within buds.

Etoposide and melphalan is an effective and well-tolerated intensive therapy regimen in advanced Hodgkin's disease. Patients in complete remission after conventional-dose salvage therapy transplanted with this regimen enjoy superior long-term DFS.

We introduce interpretable siamese neural networks (SNNs) for similarity detection to the field of theoretical physics. More precisely, we apply SNNs to events in special relativity, the transformation of electromagnetic fields, and the motion of particles in a central potential. In these examples, SNNs learn to identify data points belonging to the same event, field configuration, or trajectory of motion. We demonstrate that in the process of learning which data points belong to the same event or field configuration, these SNNs also learn the relevant symmetry invariants and conserved quantities. Such SNNs are highly interpretable, which enables us to reveal the symmetry invariants and conserved quantities without prior knowledge.

Patients with intermediate grade non-Hodgkin's lymphoma (NHL) who relapse or fail to achieve a complete remission after anthracycline-containing induction regimens have a poor outcome with conventional-dose salvage treatment. This outcome may be improved with intensive therapy and autologous transplantation (ABMT) but even in patients with proven chemotherapy-sensitive disease, relapse rates of up to 60% are observed. Reliable and powerful prognostic indicators are needed to identify appropriate patients for this expensive procedure and those subjects to whom alternative or additional treatment should be offered. We were interested in testing the hypothesis that tumour burden, and hence remission status immediately prior to transplant, is an important prognostic indicator of survival. We aggressively treated patients with conventional-dose salvage chemotherapy to maximum tumour response, and tested, by multivariate regression analysis, predictors of outcome post-transplant. We studied 81 consecutive patients with intermediate grade and immunoblastic NHL who achieved either a partial (PR) or complete remission (CR) following repetitive cycles of conventional-dose salvage therapy. Intensive therapy consisted of etoposide (60 mg/kg) and intravenous melphalan (160-180 mg/m2) with or without total body irradiation (TBI) followed by infusion of autologous unpurged bone marrow and/or blood cells. The predicted 4-year survival and progression-free survival (PFS) with a median follow-up of 37 months was 58% and 48% (95% confidence interval (CI) 37-55%), respectively. The only factor predictive of outcome was remission status at transplant (P=0.0001). The PFS at 4 years for the CR group was 61% (95% CI 53-75%). In contrast, only 25% (95% CI 11-40%) of patients undergoing autotransplant in PR were progression free at 4 years. We conclude that remission status at transplant after maximum tumour reduction is a powerful prognostic indicator.