Overcoming limitations of modelling rare species by using ensembles of small models

1. Species distribution models (SDMs) have become a standard tool in ecology and applied conservation biology. Modelling rare and threatened species is particularly important for conservation purposes. However, modelling rare species is difficult because the combination of few occurrences and many predictor variables easily leads to model overfitting. A new strategy using ensembles of small models was recently developed in an attempt to overcome this limitation of rare species modelling and has been tested successfully for only a single species so far. Here, we aim to test the approach more comprehensively on a large number of species including a transferability assessment. 2. For each species numerous small (here bivariate) models were calibrated, evaluated and averaged to an ensemble weighted by AUC scores. These 'ensembles of small models' (ESMs) were compared to standard Species Distribution Models (SDMs) using three commonly used modelling techniques (GLM, GBM, Maxent) and their ensemble prediction. We tested 107 rare and under-sampled plant species of conservation concern in Switzerland. 3. We show that ESMs performed significantly better than standard SDMs. The rarer the species, the more pronounced the effects were. ESMs were also superior to standard SDMs and their ensemble when they were independently evaluated using a transferability assessment. 4. By averaging simple small models to an ensemble, ESMs avoid overfitting without losing explanatory power through reducing the number of predictor variables. They further improve the reliability of species distribution models, especially for rare species, and thus help to overcome limitations of modelling rare species.

Rare anorectal malformation with a non-terminal colovesical fistula

We describe a unique case of anorectal malformation (ARM) with a non-terminal colovesical fistula. While some aspects are similar to the congenital pouch colon (CPC), the differences make it a distinct form.

Ventricularisation of the left atrium: a rare cause of left ventricular dysfunction in a patient with mitral valve prolapse

Case: A 11 yo girl with Marfan syndrome was referred to cardiac MR (CMR) to measure the size of her thoracic aorta. She had a typical phenotype with arachnodactyly, abnormally long arms, and was tall and slim (156 cm, 28 kg, body mass index 11,5 kg/m2). She complained of no symptoms. Cardiac auscultation revealed a prominent mid-systolic click and an end-systolic murmur at the apex. A recent echocardiogram showed a moderately dilated left ventricle with normal function and a mitral valve prolapse with moderate mitral valve regurgitation. CMR showed a dilatation of the aortic root (38 mm, Z-score 8.9) and a severe prolapse of the mitral valve with regurgitation. The ventricular cavity was moderately dilated (116 ml/m2) and its contraction was hyperdynamic (stroke volume (SV): 97 ml; LVEF 72%, with the LV volumes measured by modified Simpson method from the apex to the mitral annulus). In this patient however, the mitral prolapse was characterized by a severe backward movement of the valve toward the left atrium (LA) in systole and the dyskinetic movement of the atrioventricular plane caused a ventricularisation of a part of the LA in systole (Figure). This resulted in a significant reduction of LVEF: more than ¼ of the apparent SV was displaced backwards into the ventricularized LA volume, reducing the effective LVEF to 51% (effective SV 69ml). Moreover, by flow measurement, the SV across the ascending aorta was 30 ml (cardiac index 2.0 l/min/m2) allowing the calculation of a regurgitant fraction across the mitral valve of 56%, which was diagnostic for a severe mitral valve insufficiency. Conclusion: This case illustrates the phenomenon of a ventricularisation of the LA where the severe prolapse gives the illusion of a higher attachement of the mitral leaflets within the atrial wall. Besides the severe mitral regurgitation, this paradoxical backwards movement of the valve causes an intraventricular unloading during systole reducing the apparent LVEF of 72% to an effective LVEF of only 51%. In addition, forward flow fraction is only 22% after accounting for the regurgitant volume, as well. This combined involvement of the mitral valve could explain the discrepancy between a low output state and an apparently hyperdynamic LV contraction. Due to its ability to precisely measure flows and volumes, CMR is particularly suited to detect this phenomenon and to quantify its impact on the LV pump function.

L'infundibulite, une cause rare (?) de diabète insipide central [Infundibulitis, an unusual case of central diabetes insipidus].

We report the case of a 53 year old patient who was admitted with polyuria, polydipsia associated with fatigue, depression and sexual dysfunction. Central diabetes insipidus with hypogonadotrophic hypogonadism was diagnosed by a water restriction test and different static and dynamic hormonal dosages. Nodular thickening of the pituitary stalk was noted on the MRI and the biopsy permitted a histological diagnosis of infundibulitis.

FGFR1 and PROKR2 rare variants found in patients with combined pituitary hormone deficiencies.

The genetic aetiology of congenital hypopituitarism (CH) is not entirely elucidated. FGFR1 and PROKR2 loss-of-function mutations are classically involved in hypogonadotrophic hypogonadism (HH), however, due to the clinical and genetic overlap of HH and CH; these genes may also be involved in the pathogenesis of CH. Using a candidate gene approach, we screened 156 Brazilian patients with combined pituitary hormone deficiencies (CPHD) for loss-of-function mutations in FGFR1 and PROKR2. We identified three FGFR1 variants (p.Arg448Trp, p.Ser107Leu and p.Pro772Ser) in four unrelated patients (two males) and two PROKR2 variants (p.Arg85Cys and p.Arg248Glu) in two unrelated female patients. Five of the six patients harbouring the variants had a first-degree relative that was an unaffected carrier of it. Results of functional studies indicated that the new FGFR1 variant p.Arg448Trp is a loss-of-function variant, while p.Ser107Leu and p.Pro772Ser present signalling activity similar to the wild-type form. Regarding PROKR2 variants, results from previous functional studies indicated that p.Arg85Cys moderately compromises receptor signalling through both MAPK and Ca(2) (+) pathways while p.Arg248Glu decreases calcium mobilization but has normal MAPK activity. The presence of loss-of-function variants of FGFR1 and PROKR2 in our patients with CPHD is indicative of an adjuvant and/or modifier effect of these rare variants on the phenotype. The presence of the same variants in unaffected relatives implies that they cannot solely cause the phenotype. Other associated genetic and/or environmental modifiers may play a role in the aetiology of this condition.

Compensatory embryonic response to allele-specific inactivation of the murine X-linked gene Hcfc1.

Early in female mammalian embryonic development, cells randomly inactivate one of the two X chromosomes to achieve overall equal inactivation of parental X-linked alleles. Hcfc1 is a highly conserved X-linked mouse gene that encodes HCF-1 - a transcriptional co-regulator implicated in cell proliferation in tissue culture cells. By generating a Cre-recombinase inducible Hcfc1 knock-out (Hcfc1(lox)) allele in mice, we have probed the role of HCF-1 in actively proliferating embryonic cells and in cell-cycle re-entry of resting differentiated adult cells using a liver regeneration model. HCF-1 function is required for both extraembryonic and embryonic development. In heterozygous Hcfc1(lox/+) female embryos, however, embryonic epiblast-specific Cre-induced Hcfc1 deletion (creating an Hcfc1(epiKO) allele) around E5.5 is well tolerated; it leads to a mixture of HCF-1-positive and -negative epiblast cells owing to random X-chromosome inactivation of the wild-type or Hcfc1(epiKO) mutant allele. At E6.5 and E7.5, both HCF-1-positive and -negative epiblast cells proliferate, but gradually by E8.5, HCF-1-negative cells disappear owing to cell-cycle exit and apoptosis. Although generating a temporary developmental retardation, the loss of HCF-1-negative cells is tolerated, leading to viable heterozygous offspring with 100% skewed inactivation of the X-linked Hcfc1(epiKO) allele. In resting adult liver cells, the requirement for HCF-1 in cell proliferation was more evident as hepatocytes lacking HCF-1 fail to re-enter the cell cycle and thus to proliferate during liver regeneration. The survival of the heterozygous Hcfc1(epiKO/+) female embryos, even with half the cells genetically compromised, illustrates the developmental plasticity of the post-implantation mouse embryo - in this instance, permitting survival of females heterozygous for an X-linked embryonic lethal allele.