Early Diagnosis: End-to-End CNN-LSTM Models for Mass Spectrometry Data Classification.

Liquid chromatography-mass spectrometry (LC-MS) is a powerful method for cell profiling. The use of LC-MS technology is a tool of choice for cancer research since it provides molecular fingerprints of analyzed tissues. However, the ubiquitous presence of noise, the peaks shift between acquisitions, and the huge amount of information owing to the high dimensionality of the data make rapid and accurate cancer diagnosis a challenging task. Deep learning (DL) models are not only effective classifiers but are also well suited to jointly learn feature representation and classification tasks. This is particularly relevant when applied to raw LC-MS data and hence avoid the need for costly preprocessing and complicated feature selection. In this study, we propose a new end-to-end DL methodology that addresses all of the above challenges at once, while preserving the high potential of LC-MS data. Our DL model is designed to early discriminate between tumoral and normal tissues. It is a combination of a convolutional neural network (CNN) and a long short-term memory (LSTM) Network. The CNN network allows for significantly reducing the high dimensionality of the data while learning spatially relevant features. The LSTM network enables our model to capture temporal patterns. We show that our model outperforms not only benchmark models but also state-of-the-art models developed on the same data. Our framework is a promising strategy for improving early cancer detection during a diagnostic process.

Cumulative learning enables convolutional neural network representations for small mass spectrometry data classification.

Rapid and accurate clinical diagnosis remains challenging. A component of diagnosis tool development is the design of effective classification models with Mass spectrometry (MS) data. Some Machine Learning approaches have been investigated but these models require time-consuming preprocessing steps to remove artifacts, making them unsuitable for rapid analysis. Convolutional Neural Networks (CNNs) have been found to perform well under such circumstances since they can learn representations from raw data. However, their effectiveness decreases when the number of available training samples is small, which is a common situation in medicine. In this work, we investigate transfer learning on 1D-CNNs, then we develop a cumulative learning method when transfer learning is not powerful enough. We propose to train the same model through several classification tasks over various small datasets to accumulate knowledge in the resulting representation. By using rat brain as the initial training dataset, a cumulative learning approach can have a classification accuracy exceeding 98% for 1D clinical MS-data. We show the use of cumulative learning using datasets generated in different biological contexts, on different organisms, and acquired by different instruments. Here we show a promising strategy for improving MS data classification accuracy when only small numbers of samples are available.

Interplay between Jasmonic Acid, Phosphate Signaling and the Regulation of Glycerolipid Homeostasis in Arabidopsis.

Jasmonic acid (JA) biosynthesis and signaling are activated in Arabidopsis cultivated in phosphate (Pi) deprived conditions. This activation occurs mainly in photosynthetic tissues and is less important in roots. In leaves, the enhanced biosynthesis of JA coincides with membrane glycerolipid remodeling triggered by the lack of Pi. We addressed the possible role of JA on the dynamics and magnitude of glycerolipid remodeling in response to Pi deprivation and resupply. Based on combined analyses of gene expression, JA biosynthesis and glycerolipid remodeling in wild-type Arabidopsis and in the coi1-16 mutant, JA signaling seems important in the determination of the basal levels of phosphatidylcholine, phosphatidic acid (PA), monogalactosyldiacylglycerol (MGDG) and digalactosyldiacylglycerol. JA impact on MGDG steady state level and fluctuations seem contradictory. In the coi1-16 mutant, the steady state level of MGDG is higher, possibly due to a higher level of PA in the mutant, activating MGD1, and to an increased expression of MGD3. These results support a possible impact of JA in limiting the overall content of this lipid. Concerning lipid variations, upon Pi deprivation, JA seems rather associated with a specific MGDG increase. Following Pi resupply, whereas the expression of glycerolipid remodeling genes returns to basal level, JA biosynthesis and signaling genes are still upregulated, likely due to a JA-induced positive feedback remaining active. Distinct impacts on enzymes synthesizing MGDG, that is, downregulating MGD3, possibly activating MGD1 expression and limiting the activation of MGD1 via PA, might allow JA playing a role in a sophisticated fine tuning of galactolipid variations.

Screening for Biologically Annotated Drugs That Trigger Triacylglycerol Accumulation in the Diatom Phaeodactylum.

Microalgae are a promising feedstock for the production of triacylglycerol (TAG) for a variety of potential applications, ranging from food and human health to biofuels and green chemistry. However, obtaining high TAG yields is challenging. A phenotypic assay for the accumulation of oil droplets was developed to screen a library of 1,200 drugs, annotated with pharmacology information, to select compounds that trigger TAG accumulation in the diatom Phaeodactylum tricornutum Using this screen, we identified 34 molecules acting in a dose-dependent manner. Previously characterized targets of these compounds include cell division and cell signaling effectors, membrane receptors and transporters, and sterol metabolism. Among the five compounds possibly acting on sterol metabolism, we focused our study on ethynylestradiol, a synthetic form of estrogen that is used in contraceptive pills and known for its ecological impact as an endocrine disruptor. Ethynylestradiol impaired the production of very-long-chain polyunsaturated fatty acids, destabilized the galactolipid versus phospholipid balance, and triggered the recycling of fatty acids from membrane lipids to TAG. The P. tricornutum transcriptomic response to treatment with ethynylestradiol was consistent with the reallocation of carbon from sterols to acetyl-coenzyme A and TAG. The mode of action and catabolism of ethynylestradiol are unknown but might involve several up-regulated cytochrome P450 proteins. A fatty acid elongase, Δ6-ELO-B1, might be involved in the impairment of very-long-chain polyunsaturated fatty acids and fatty acid turnover. This phenotypic screen opens new perspectives for the exploration of novel bioactive molecules, potential target genes, and pathways controlling TAG biosynthesis. It also unraveled the sensitivity of diatoms to endocrine disruptors, highlighting an impact of anthropogenic pollution on phytoplankton.

Sequencing, De Novo Assembly, and Annotation of the Complete Genome of a New Thraustochytrid Species, Strain CCAP_4062/3.

Thraustochytrids are ecologically and biotechnologically relevant marine species. We report here the de novo assembly and annotation of the whole-genome sequence of a new thraustochytrid strain, CCAP_4062/3. The genome size was estimated at 38.7 Mb with 11,853 predicted coding sequences, and the GC content was scored at 57%.

Nitric Oxide Mediates Nitrite-Sensing and Acclimation and Triggers a Remodeling of Lipids.

Nitric oxide (NO) is an intermediate of the nitrogen cycle, an industrial pollutant, and a marker of climate change. NO also acts as a gaseous transmitter in a variety of biological processes. The impact of environmental NO needs to be addressed. In diatoms, a dominant phylum in phytoplankton, NO was reported to mediate programmed cell death in response to diatom-derived polyunsaturated aldehydes. Here, using the Phaeodactylum Pt1 strain, 2E,4E-decadienal supplied in the micromolar concentration range led to a nonspecific cell toxicity. We reexamined NO biosynthesis and response in Phaeodactylum NO inhibits cell growth and triggers triacylglycerol (TAG) accumulation. Feeding experiments indicate that NO is not produced from Arg but via conversion of nitrite by the nitrate reductase. Genome-wide transcriptional analysis shows that NO up-regulates the expression of the plastid nitrite reductase and genes involved in the subsequent incorporation of ammonium into amino acids, via both Gln synthesis and Orn-urea pathway. The phosphoenolpyruvate dehydrogenase complex is also up-regulated, leading to the production of acetyl-CoA, which can feed TAG accumulation upon exposure to NO. Transcriptional reprogramming leading to higher TAG content is balanced with a decrease of monogalactosyldiacylglycerol (MGDG) in the plastid via posttranslational inhibition of MGDG synthase enzymatic activity by NO. Intracellular and transient NO emission acts therefore at the basis of a nitrite-sensing and acclimating system, whereas a long exposure to NO can additionally induce a redirection of carbon to neutral lipids and a stress response.