Extracting diagnoses and investigation results from unstructured text in electronic health records by semi-supervised machine learning

Background

Electronic health records are invaluable for medical research, but much of the information is recorded as unstructured free text which is time-consuming to review manually.

Aim

To develop an algorithm to identify relevant free texts automatically based on labelled examples.

Methods

We developed a novel machine learning algorithm, the ‘Semi-supervised Set Covering Machine’ (S3CM), and tested its ability to detect the presence of coronary angiogram results and ovarian cancer diagnoses in free text in the General Practice Research Database. For training the algorithm, we used texts classified as positive and negative according to their associated Read diagnostic codes, rather than by manual annotation. We evaluated the precision (positive predictive value) and recall (sensitivity) of S3CM in classifying unlabelled texts against the gold standard of manual review. We compared the performance of S3CM with the Transductive Vector Support Machine (TVSM), the original fully-supervised Set Covering Machine (SCM) and our ‘Freetext Matching Algorithm’ natural language processor.

Results

Only 60% of texts with Read codes for angiogram actually contained angiogram results. However, the S3CM algorithm achieved 87% recall with 64% precision on detecting coronary angiogram results, outperforming the fully-supervised SCM (recall 78%, precision 60%) and TSVM (recall 2%, precision 3%). For ovarian cancer diagnoses, S3CM had higher recall than the other algorithms tested (86%). The Freetext Matching Algorithm had better precision than S3CM (85% versus 74%) but lower recall (62%).

Conclusions

Our novel S3CM machine learning algorithm effectively detected free texts in primary care records associated with angiogram results and ovarian cancer diagnoses, after training on pre-classified test sets. It should be easy to adapt to other disease areas as it does not rely on linguistic rules, but needs further testing in other electronic health record datasets.     

Flour carbohydrate catabolism and metabolite production by sourdough lactic acid bacteria

The aim of this study was to assess the mode of carbohydrate catabolism by lactic acid bacteria isolated from traditional sourdoughs, as well as to study their effect on the metabolites produced. For this purpose, single cultures of the heterofermentative lactic acid bacteria Lactobacillus sanfranciscensis, Lactobacillus brevis, Weissella cibaria, and the homofermentative Lactobacillus paralimentarius and Pediococcus pentosaceus were grown in liquid media containing glucose, fructose, maltose and sucrose, either as a single carbon source or in combination with glucose. Carbon catabolism and the production of metabolites were determined by HPLC analysis. W. cibaria could ferment all carbon sources, L. sanfranciscensis, L. paralimentarius and P. pentosaceus could not ferment sucrose, while L. brevis could only ferment maltose. The presence of glucose did not influence the utilization of fructose and maltose by L. sanfranciscensis, while it repressed the fermentation of fructose, maltose and sucrose by W. cibaria, and fructose and maltose by L. paralimentarius and P. pentosaceus. Moreover, L. sanfranciscensis and L. brevis could obtain extra ATP through the reduction of fructose to mannitol, which favored the production of acetic acid against ethanol. The utilization of fructose as an electron acceptor has a decisive effect on the prevailing of L. sanfranciscensis and L. brevis in spontaneously fermented sourdough and in the scarce appearance of the other lactic acid bacteria studied.     

三裂叶野葛毛状根的培养及其葛根素的产生

发根农杆菌ATCC15834感染三裂叶野葛叶片外植体20天后,从其切口叶脉处产生的愈伤组织上产生毛状根。感染35天后约85%的叶片外植体产生毛状根。毛状根能在无外源生长调节剂的MS固体和液体培养基上自主生长,但在液体培养基中培养的毛状根生长更迅速,也不会形成愈伤组织。毛状根线粒体膜电势的荧光染色结果表明,液体培养的毛状根细胞线粒体的膜电势比固体培养的毛状根高11.8倍。PCR结果证实,发根农杆菌Ri质粒的rolB和rolC基因已在三裂叶野葛毛状根基因组中整合并得到表达。HPLC测定结果表明,三裂叶野葛毛状根中的葛根素含量约为对照根(种子萌发产生的幼苗根)的2.5倍,达1.190 mg/g.dry.wt;并比多年生葛根生药片的葛根素含量高6.7%。     

Engineering stem cell fate with biochemical and biomechanical properties of microcarriers

Microcarriers have been widely used for various biotechnology applications because of their high scale‐up potential, high reproducibility in regulating cellular behavior, and well‐documented compliance with current Good Manufacturing Practices (cGMP). Recently, microcarriers have been emerging as a novel approach for stem cell expansion and differentiation, enabling potential scale‐up of stem cell‐derived products in large bioreactors. This review summarizes recent advances of using microcarriers in mesenchymal stem cell (MSC) and pluripotent stem cell (PSC) cultures. From the reported data, efficient expansion and differentiation of stem cells on microcarriers rely on their ability to modulate cell shape (i.e. round or spreading) and cell organization (i.e. aggregate size). Nonetheless, current screening of microcarriers remains empirical, and accurate understanding of how stem cells interact with microcarriers still remains unknown. This review suggests that accurate characterization of biochemical and biomechanical properties of microcarriers is required to fully exploit their potential in regulating stem cell fate decision. Due to the variety of microcarriers, such detailed analyses should lead to the rational design of application‐specific microcarriers, enabling the exploitation of reproducible effects for large scale biomedical applications. © 2013 American Institute of Chemical Engineers Biotechnol. Prog., 29:1354–1366, 2013     

Mitonuclear discordance as a confounding factor in the DNA taxonomy of monogonont rotifers

Discordance between mitochondrial and nuclear phylogenies is being increasingly recognized in animals and may confound DNA‐based taxonomy. This is especially relevant for taxa whose microscopic size often challenges any effort to distinguish between cryptic species without the assistance of molecular data. Regarding mitonuclear discordance, two strikingly contrasting scenarios have been recently demonstrated in the monogonont rotifers of the genus Brachionus. While strict mitonuclear concordance was observed in the marine B. plicatilis species complex, widespread hybridization‐driven mitonuclear discordance was revealed in the freshwater B. calyciflorus species complex. Here, we investigated the frequency of occurrence and the potential drivers of mitonuclear discordance in three additional freshwater monogonont rotifer taxa, and assessed its potential impact on the reliability of DNA taxonomy results based on commonly used single markers. We studied the cryptic species complexes of Keratella cochlearis, Polyarthra dolichoptera and Synchaeta pectinata. Phylogenetic reconstructions were based on the mitochondrial barcoding marker cytochrome c oxidase subunit I gene and the nuclear internal transcribed spacer 1 locus, which currently represent the two most typical genetic markers used in rotifer DNA taxonomy. Species were delimited according to each marker separately using a combination of tree‐based coalescent, distance‐based and allele‐sharing‐based approaches. Mitonuclear discordance was observed in all species complexes with incomplete lineage sorting and unresolved phylogenetic reconstructions recognized as the likely drivers. Evidence from additional sources, such as morphology and ecology, is thus advisable for deciding between often contrasting mitochondrial and nuclear species scenarios in these organisms.     

Pigmented variants of Tolypocladium inflatum in relation to cyclosporin A production

Summary We have isolated four distinct colony types of the fungus Tolypocladium inflatum, the producer of the immunosuppresive agent cyclosporin A: morphologically normal white, red, and orange colonies and morphologically diverse tiny brown colonies. In liquid cultures, white normal and brown colonies developed into yellow broths. The broth of the brown colony had a low final pH and low cyclosporin production, whereas orange and red colonies had dark brown and even black broths with higher final pH and high cyclosporin production. The specific production of cyclosporin A by the red colony was three times that of the white normal colonies. Offprint requests to: S. N. Agathos     

Monogonont rotifers as model systems for the study of micro-evolutionary adaptation and its eco-evolutionary implications

A better understanding of the ability of organisms to adapt to local selection conditions is essential for a better insight in their ecological dynamics. The study of micro-evolutionary adaptation and its eco-evolutionary consequences is challenging for many reasons and the choice of a suitable model organism is particularly important. In this paper, we explain why monogonont rotifers, through their unique combination of traits, are ideal study organisms for this purpose. With a literature review, we demonstrate the capacity of monogonont populations to adapt to a variety of selection conditions (e.g., salinity, food shortage, elemental limitation, and disturbance regimes) within very short-time frames and highlight some potential eco-evolutionary implications. Although monogononts are increasingly used in eco-evolution-oriented studies, their potential is still underappreciated compared to other model organisms. No doubt the high prevalence of cryptic species complexes and the lack of genomic tools form important obstacles that may discourage researchers to work with this group. Here, we argue that none of these difficulties should prevent monogonont rotifers from becoming commonly used model organisms in micro-evolutionary studies and make suggestions for future research.