Biology subjects

Campagne, F.

Publications and source records attributed to Campagne, F..

5 recordsLinked to original sources

bioRxiv2018

GenotypeTensors: Efficient Neural Network Genotype Callers

We studied the problem of calling genotypes using neural networks. A machine learning approach to calling genotypes requires a training set, an approach to convert genomic sites into tensors and robust model development and evaluation protocols. We discuss each of these components of our approach and compare four types of neural network training protocols, two fully supervised and two semi-supervised approaches. Semi-supervised approaches use unlabeled data to supplement limited quantities of labeled data. Random hyper-parameter searches identified highly performing models that reach indel F1 of 99.4% on a chromosomes 20, 21, 22 and X of NA12878/HG001. We further validate these models by evaluating performance on HG002, an independent sample used in the PrecisionFDA challenge. We apply GenotypeTensors to evaluate the impact of (1) training with small datasets, (2) training models only with sites inside confidence regions, or (3) training with improved true label annotations. A PyTorch open-source implementation of GenotypeTensors is available at https://github.com/CampagneLaboratory/GenotypeTensors. DNANexus cloud applications are provided to help process new datasets both to train model or call genotypes with trained models.

bioinformatics↗

bioRxiv2017

Tacrolimus Rescues Endothelial ALK1 Loss-Of-Function Signaling And Improves HHT Vascular Pathology

Hereditary hemorrhagic telangiectasia (HHT) is a genetic vascular disorder arising from endothelial cell (EC) proliferation and hypervascularization, for which no cure exists. Because HHT is caused by loss-of-function mutations in BMP9-ALK1-Smad1/5/8 signaling, interventions aimed at activating this pathway are of therapeutic value. By screening FDA-approved drug libraries, we identified tacrolimus (FK-506) as a potent activator of Smad1/5/8 in BMP9-challenged reporter cells. In primary ECs, tacrolimus activated Smad1/5/8 to oppose the pro-angiogenic gene expression signature associated with ALK1 loss-of-function, by notably reducing Dll4 expression. In these cells, tacrolimus also inhibited Akt and p38 stimulation by VEGF. In the BMP9/10-immunodepleted postnatal retina--a mouse model of HHT vascular pathology--tacrolimus activated endothelial Smad1/5/8 and prevented the Dll4 overexpression and hypervascularization associated with this model. Finally, tacrolimus stimulated Smad1/5/8 in cells transfected with BMP9-unresponsive ALK1 HHT mutants and in HHT patient blood outgrowth ECs. We propose that tacrolimus repurposing has therapeutic potential in HHT.

cell biology↗

bioRxiv2016

Training Genotype Callers with Neural Networks

We present an open source software toolkit for training deep learning models to call genotypes in high-throughput sequencing data. The software supports SAM, BAM, CRAM and Goby alignments and the training of models for a variety of experimental assays and analysis protocols. We evaluate this software in the Illumina Platinum whole genome datasets and find that a deep learning model trained on 80% of the genome achieves a 0.986% accuracy on variants (genotype concordance) when trained with 10% of the data from a genome. The software is distributed at https://github.com/CampagneLaboratory/variationanalysis. The software makes it possible to train genotype calling models on consumer hardware with CPUs or GPU(s). It will enable individual investigators and small laboratories to train and evaluate their own models and to make open source contributions. We welcome contributions to extend this early prototype or evaluate its performance on other gold standard datasets.

bioinformatics↗

bioRxiv2016

CONTINUATION: Evaluation of adaptive somatic models in a gold standard whole genome somatic dataset

In http://dx.doi.org/10.1101/079087, we presented adaptive models for calling somatic mutations in high-throughput sequencing data. These models were developed by training deep neural networks with semi-simulated data. In this continuation, I evaluate how such models can predict known somatic mutations in a real dataset. To address this question, I tested the approach using samples from the International Cancer Genome Consortium (ICGC) and the previously published ground-truth mutations (GoldSet). This evaluation revealed that training models with semi-simulation does produce models that exhibit strong performance in real datasets. I found a linear relationship between the performance observed on a semi-simulated validation set and independent ground-truth in the gold set (R2 = 0.952, P < 2-16). I also found that semi-simulation can be used to pre-train models before continuing training with true labels and that this pre-training improves model performance substantially on the real dataset compared to training models only with the real dataset. The best model pre-trained with semi-simulation achieved an AUC of 0.969 [0.957-0.982] (95% confidence interval) compared to 0.911 [0.890-0.932] when training with real labels only. These data demonstrate that semi-simulation can be a very effective approach to training filtering and ranking probabilistic models.

bioinformatics↗

bioRxiv2016

A mouse model of hereditary hemorrhagic telangiectasia generated by transmammary-delivered immunoblocking of BMP9 and BMP10

Hereditary hemorrhagic telangiectasia (HHT) is a potentially life-threatening genetic vascular disorder caused by loss-of-function mutations in the genes encoding activin receptor-like kinase 1 (ALK1), endoglin, Smad4, and bone morphogenetic protein 9 (BMP9). Injections of mouse neonates with BMP9/10 blocking antibodies lead to HHT-like vascular defects in the postnatal retinal angiogenesis model. Mothers and newborns share the same immunity through the transfer of maternal antibodies during breastfeeding. Here, we investigated whether the transmammary delivery route could improve the ease and consistency of administering anti-BMP9/10 antibodies in the postnatal retinal angiogenesis model. We found that anti-BMP9/10 antibodies, when intraperitoneally injected into lactating dams, are efficiently transferred into the circulation of breastfed neonatal pups. Strikingly, pups receiving anti-BMP9/10 antibodies via breastfeeding displayed consistent and robust vascular pathology in the retina, which included hypervascularization and defects in arteriovenous specification, as well as the presence of multiple and massive arteriovenous malformations. Furthermore, RNA-Seq analyses of neonatal retinas identified an increase in the key pro-angiogenic factor, angiopoietin-2, as the most significant change in gene expression triggered by the transmammary delivery of anti-BMP9/10 antibodies. Transmammary-delivered BMP9/10 immunoblocking in the mouse neonatal retina is therefore a practical, noninvasive, reliable, and robust model of HHT vascular pathology.

cell biology↗