Machine learning-based reclassification of germline variants of unknown significance: The RENOVO algorithm.

The increasing scope of genetic testing allowed by next-generation sequencing (NGS) dramatically increased the number of genetic variants to be interpreted as pathogenic or benign for adequate patient management. Still, the interpretation process often fails to deliver a clear classification, resulting in either variants of unknown significance (VUSs) or variants with conflicting interpretation of pathogenicity (CIP); these represent a major clinical problem because they do not provide useful information for decision-making, causing a large fraction of genetically determined disease to remain undertreated. We developed a machine learning (random forest)-based tool, RENOVO, that classifies variants as pathogenic or benign on the basis of publicly available information and provides a pathogenicity likelihood score (PLS). Using the same feature classes recommended by guidelines, we trained RENOVO on established pathogenic/benign variants in ClinVar (training set accuracy = 99%) and tested its performance on variants whose interpretation has changed over time (test set accuracy = 95%). We further validated the algorithm on additional datasets including unreported variants validated either through expert consensus (ENIGMA) or laboratory-based functional techniques (on BRCA1/2 and SCN5A). On all datasets, RENOVO outperformed existing automated interpretation tools. On the basis of the above validation metrics, we assigned a defined PLS to all existing ClinVar VUSs, proposing a reclassification for 67% with >90% estimated precision. RENOVO provides a validated tool to reduce the fraction of uninterpreted or misinterpreted variants, tackling an area of unmet need in modern clinical genetics.

Spatial-positional associations in short-term memory can vanish in long-term memory.

Studies on the SPoARC effect have shown that serial information is spatially processed in working memory. However, it remains unknown whether these spatial-positional associations are durable or only temporary. This study aimed at investigating whether spatialization would persist when a sequence presented repeatedly is expected to be chunked. If chunked, the items could be unified spatially and their spatialization could vanish. Thirty-seven participants performed a spatialization task which was remotely inspired by the Hebb repetition paradigm. A sequence of four stimuli presented individually in the middle of a computer screen was repeated throughout the task. After each sequence, participants had to decide whether a probe belonged to the series using two lateralized response keys. The results showed no spatialization for these repetitive sequences, on average. Moreover, further analysis revealed that the effect was detectable at the beginning of the task, suggesting that the more the sequence was repeated, the less participants spatialized information from left to right. These findings show that associations created in working memory between items and space can vanish in repeated sequences: we discuss the idea that working memory progressively saves on spatialization once a sequence is chunked in long-term memory.

TMBleR: a bioinformatic tool to optimize TMB estimation and predictive power.

MOTIVATION: Tumor mutational burden (TMB) has been proposed as a predictive biomarker for immunotherapy response in cancer patients, as it is thought to enrich for tumors with high neoantigen load. TMB assessed by whole-exome sequencing is considered the gold standard but remains confined to research settings. In the clinical setting, targeted gene panels sampling various genomic sizes along with diverse strategies to estimate TMB were proposed and no real standard has emerged yet. RESULTS: We provide the community with TMBleR, a tool to measure the clinical impact of various strategies of panel-based TMB measurement. AVAILABILITY AND IMPLEMENTATION: R package and docker container (GPL-3 Open Source license): https://acc-bioinfo.github.io/TMBleR/. Graphical-user interface website: https://bioserver.ieo.it/shiny/app/tmbler. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Bootstrapping the visuospatial bootstrapping effect and testing its spatialisation.

When sequences of digits are visually presented within a numerical keypad on a screen, memory span increases, this effect was named visuospatial bootstrapping. The aim of the first experiment was to know if this effect could emerge without presenting a keypad on the screen. For this purpose, a three-phase experiment was designed. During phase 1, the immediate serial recall of two groups of participants was compared (pre-training): the first group saw sequences of one-digit numbers displayed on a screen within a keypad (the keypad group) whereas the second group heard the (same) sequences (the auditive group). During phase 2, all participants underwent a training session to help them visualise in their mind a keypad. Finally, in the third phase, participants were tested again with an immediate serial recall task (post-training). Results showed that both groups had comparable performance in post-training indicating that the visuospatial bootstrapping could be obtained without displaying a numerical keypad. The second experiment also involved a keypad group and an auditive group and was designed to investigate their spatial representation. Results showed that both groups spatialised the digits following the keypad spatial configuration: digits 1-4-7 were associated to left, 2-5-8 to middle and 3-6-9 to right.

Can spatialisation be extended to episodic memory and open sets?

The SPoARC effect (Spatial Positional Associated Response Codes) has only been observed in working memory (WM) using closed sets. It is interpreted as showing that individuals spatialise to-be-remembered items in a left-to-right fashion, using spatialisation as context. Given that context is crucial for episodic memory (EM), we tested if this effect could be observed in EM by using 15-word lists taken from an open set. After each list, 30 probes were sequentially displayed to test recognition. The left/right-hand key assignment for yes/no answers was varied. No SPoARC effect was observed. However, as all previous SPoARC experiments had used short lists and closed sets, it was not possible to know if this absence of SPoARC was due to the open set feature or the length of the lists. A second experiment was thus run using open sets and short 5-word lists, which do not necessitate EM to be remembered. A SPoARC effect was observed indicating that Experiment 1 result was due to the involvement of supra-span lists and that SPoARC effects do not extend to EM with open sets. Experiment 2 also enabled us to generalise the SPoARC effect to open sets in WM for the first time.

Distinctiveness as a function of spatial expansion in verbal working memory: comment on Kreitz, Furley, Memmert, and Simons (2015).

In a recent study, Kreitz et al. (Psychological Research 79:1034-1041, 2015) reported on a relationship between verbal working memory capacity and visuo-spatial attentional breadth. The authors hinted at attentional control to be the major link underlying this relationship. We put forward an alternative explanation by framing it within the context of a recent theory on serial order in memory: verbal item sequences entering in working memory are coded by adding a spatial context that can be derived from reading/writing habits. The observation by Kreitz et al. (Psychological Research 79:1034-1041, 2015) enriches this framework by suggesting that a larger visuo-spatial attentional breadth allows for internal coding of the verbal items in a more (spatially) distinct manner-thereby increasing working memory performance. As such, Kreitz et al. (Psychological Research 79:1034-1041, 2015) is the first study revealing a functional link between visuo-spatial attentional breadth and verbal working memory size, which strengthens spatial accounts of serial order coding in working memory.

Promoting the experimental dialogue between working memory and chunking: Behavioral data and simulation.

Working memory (WM) is a cognitive system allowing short-term maintenance and processing of information. Maintaining information in WM consists, classically, in rehearsing or refreshing it. Chunking could also be considered as a maintenance mechanism. However, in the literature, it is more often used to explain performance than explicitly investigated within WM paradigms. Hence, the aim of the present paper was (1) to strengthen the experimental dialogue between WM and chunking, by studying the effect of acronyms in a computer-paced complex span task paradigm and (2) to formalize explicitly this dialogue within a computational model. Young adults performed a WM complex span task in which they had to maintain series of 7 letters for further recall while performing a concurrent location judgment task. The series to be remembered were either random strings of letters or strings containing a 3-letter acronym that appeared in position 1, 3, or 5 in the series. Together, the data and simulations provide a better understanding of the maintenance mechanisms taking place in WM and its interplay with long-term memory. Indeed, the behavioral WM performance lends evidence to the functional characteristics of chunking that seems to be, especially in a WM complex span task, an attentional time-based mechanism that certainly enhances WM performance but also competes with other processes at hand in WM. Computational simulations support and delineate such a conception by showing that searching for a chunk in long-term memory involves attentionally demanding subprocesses that essentially take place during the encoding phases of the task.

Becoming an expert: Ontogeny of expertise as an example of neural reuse.

In this commentary, we discuss an important pattern of results in the literature on the neural basis of expertise: (a) decrease of cerebral activation at the beginning of acquisition of expertise and (b) functional cerebral reorganization as a consequence of years of practice. We show how these two results can be integrated with the neural reuse framework.

Hearing the inaudible experimental subject: Echoes of Inaudi, Binet's calculating prodigy.

Historians of psychology have traditionally focused on ideas (intellectual history), the "great men" who produced them (an older style of biography sometimes called "hagiography"), or-more recently-the influence of the contexts that shaped them (social and cultural history). A still more recent approach is to bring in those invisible subjects whose experiences have previously been ignored, most often through histories focusing on the discipline's forgotten women or minority contributors: "history from below" (subaltern history). A variation on this was popularized in the history of psychiatry (viz., "patient voices") and has since been carried into the history of psychology (e.g., "feminist voices"). The latest innovation is to focus on what Jill Morawski has referred to as "the discipline's experimental subjects." (These are the collective done-to, rather than the doers, of psychological research.) This history is one of those: an attempt to look behind Alfred Binet to find an influence that shaped his work. The purpose is thus to "give voice" to this unheard-from subject-the until-now inaudible Jacques Inaudi (including excerpts from newspaper interviews and translations from his recently discovered autobiography)-and at the same time advance Morawski's historiographical project. We then get a glimpse of what it was like to be a child prodigy in France in the 1880s, as well as what securing scientific patrons could do for one's prospects. By focusing specifically on Binet's unheard-from experimental subject, we are also afforded new perspectives of the history of late-19th century French psychology (reflecting another emerging interest, "international history"), and we gain new insights into the prehistory of contemporary Binet-style intelligence testing.

EXPRESS:Spatial Organisation in the Human Mind as a Function of the Distance Between Stimuli.

Studies investigating serial order in working memory have shown that participants from Western cultures are faster at responding to items presented at the beginning of a sequence using their left hand and faster at responding to items at the end with their right hand. This is known as the spatial positional association of response codes (SPoARC) effect. The SPoARC effect provides evidence that recently presented information is spatially organised in the cognitive system along a horizontal axis. This study investigated the flexibility of spatialisation by testing the effect that distance between items presented on a screen has on the magnitude of the SPoARC effect. It was hypothesised that by increasing the distance between items on a screen a larger SPoARC effect would be found. We used three conditions: central, narrow, and wide. In central, four random letters were presented sequentially at the centre of the screen, in narrow the letters were presented from left to right on the screen, wide was the same as narrow but the separation between the letters was larger. Participants consisted of 64 adults aged 18-55 years old. Participants were presented with four random letters, followed by single probe letter, participants had to indicate, by pressing a key on a normal keyboard, if the probe had been in the sequence. We analysed the data with multilevel modelling. We found evidence for the SPoARC effect in all three conditions. But no evidence that the effect varied between conditions.

The effect of long-term working memory through personalization applied to free recall: uncurbing the primacy-effect enthusiasm.

In this study, a personalization method (Guida, Tardieu, & Nicolas, European Journal of Cognitive Psychology, 21: 862-896 2009) was applied to a free-recall task. Fifteen pairs of words, composed of an object and a location, were presented to 93 participants, who had to mentally associate each pair and subsequently recall the objects. A 30-s delay was introduced on half of the trials, the presentation rate was manipulated (5 or 10 s per item), and verbal and visuospatial working memory tests were administered to test for their effects on the serial curve. Two groups were constituted: a personalized group, for whom the locations were well-known places on their university campus, and a nonpersonalized group, for whom the locations did not refer to known places. Since personalization putatively operationalizes long-term working memory (Ericsson & Kintsch, Psychological Review, 102: 211-245 1995)-namely, the capacity to store information reliably and rapidly in long-term memory-and if we take a dual-store approach to memory, the personalization advantage would be expected to be greater for pre-recency than for recency items. Overall, the results were compatible with long-term working memory theory. They contribute to validating the personalization method as a methodology to characterize the contribution of long-term memory storage to performance in working memory tasks.

Optimized experimental designs to best detect spatial positional association of response codes in working memory.

The SPoARC (Spatial Positional Association of Response Codes) effect refers to spatialization of information in working memory. Among the potential factors that could influence how order is mapped onto a mental space during the recognition process, we selected the following two factors: i) the type of stimuli, in particular their verbal vs. visual aspects and ii) the number of probes. In this study, 137 participants memorized sequences of either words or pictures and subsequently performed a recognition test for which they responded using lateralized keys. For half of the participants, only one probe was presented after each sequence, whereas the other half was administered several probes. A significantly greater number of participants presented a SPoARC using a single probe. We discuss that spatialization is best detected when the sequence is scanned only once. Results also showed no difference between the two types of stimuli (i.e., verbal vs. visual). This finding raises the question of the respective roles of verbalization and visualization in the SPoARC.

How chunks, long-term working memory and templates offer a cognitive explanation for neuroimaging data on expertise acquisition: a two-stage framework.

Our review of research on PET and fMRI neuroimaging of experts and expertise acquisition reveals two apparently discordant patterns in working-memory-related tasks. When experts are involved, studies show activations in brain regions typically activated during long-term memory tasks that are not observed with novices, a result that is compatible with functional brain reorganization. By contrast, when involving novices and training programs, studies show a decrease in brain regions typically activated during working memory tasks, with no functional reorganization. We suggest that the latter result is a consequence of practice periods that do not allow important structures to be completely acquired: knowledge structures (i.e., Ericsson and Kintsch's retrieval structures; Gobet and Simon's templates) and in a lesser way, chunks. These structures allow individuals to improve performance on working-memory tasks, by enabling them to use part of long-term memory as working memory, causing a cerebral functional reorganization. Our hypothesis is that the two brain activation patterns observed in the literature are not discordant, but involve the same process of expertise acquisition in two stages: from decreased activation to brain functional reorganization. The dynamic of these two physiological stages depend on the two above-mentioned psychological constructs: chunks and knowledge structures.

A SPoARC of Music: Musicians Spatialize Melodies but not All-Comers.

Recent studies on the spatial positional associated response codes (SPoARC) effect have shown that when Western adults are asked to keep in mind sequences of verbal items, they mentally spatialize them along the horizontal axis, with the initial items being associated with the left and the last items being associated with the right. The origin of this mental line is still debated, but it has been theorized that it necessitates specific spatial cognitive structures to emerge, which are built through expertise. This hypothesis is examined by testing for the first time whether Western individuals spatialize melodies from left to right and whether expertise in the musical domain is necessary for this effect to emerge. Two groups (musicians and non-musicians) of participants were asked to memorize sequences of four musical notes and to indicate if a subsequent probe was part of the sequence by pressing a "yes" key or a "no" key with the left or right index finger. Left/right-hand key assignment was reversed at mid-experiment. The results showed a SPoARC effect only for the group of musicians. Moreover, no association between pitch and hand responses was observed in either of the two groups. These findings suggest a crucial role of expertise in the SPoARC effect.

Using RNA-seq to determine the transcriptional landscape and the hypoxic response of the pathogenic yeast Candida parapsilosis.

BACKGROUND: Candida parapsilosis is one of the most common causes of Candida infection worldwide. However, the genome sequence annotation was made without experimental validation and little is known about the transcriptional landscape. The transcriptional response of C. parapsilosis to hypoxic (low oxygen) conditions, such as those encountered in the host, is also relatively unexplored. RESULTS: We used next generation sequencing (RNA-seq) to determine the transcriptional profile of C. parapsilosis growing in several conditions including different media, temperatures and oxygen concentrations. We identified 395 novel protein-coding sequences that had not previously been annotated. We removed > 300 unsupported gene models, and corrected approximately 900. We mapped the 5' and 3' UTR for thousands of genes. We also identified 422 introns, including two introns in the 3' UTR of one gene. This is the first report of 3' UTR introns in the Saccharomycotina. Comparing the introns in coding sequences with other species shows that small numbers have been gained and lost throughout evolution. Our analysis also identified a number of novel transcriptional active regions (nTARs). We used both RNA-seq and microarray analysis to determine the transcriptional profile of cells grown in normoxic and hypoxic conditions in rich media, and we showed that there was a high correlation between the approaches. We also generated a knockout of the UPC2 transcriptional regulator, and we found that similar to C. albicans, Upc2 is required for conferring resistance to azole drugs, and for regulation of expression of the ergosterol pathway in hypoxia. CONCLUSION: We provide the first detailed annotation of the C. parapsilosis genome, based on gene predictions and transcriptional analysis. We identified a number of novel ORFs and other transcribed regions, and detected transcripts from approximately 90% of the annotated protein coding genes. We found that the transcription factor Upc2 role has a conserved role as a major regulator of the hypoxic response in C. parapsilosis and C. albicans.

Regulation of the hypoxic response in Candida albicans.

The regulation of the response of Candida albicans to hypoxic (low-oxygen) conditions is poorly understood. We used microarray and other transcriptional analyses to investigate the role of the Upc2 and Bcr1 transcription factors in controlling expression of genes involved in cell wall metabolism, ergosterol synthesis, and glycolysis during adaptation to hypoxia. Hypoxic induction of the ergosterol pathway is mimicked by treatment with sterol-lowering drugs (ketoconazole) and requires UPC2. Expression of three members of the family CFEM (common in several fungal extracellular membranes) of cell wall genes (RBT5, PGA7, and PGA10) is also induced by hypoxia and ketoconazole and requires both UPC2 and BCR1. Expression of glycolytic genes is induced by hypoxia but not by treatment with sterol-lowering drugs, whereas expression of respiratory pathway genes is repressed. However, Upc2 does not play a major role in regulating expression of genes required for central carbon metabolism. Our results indicate that regulation of gene expression in response to hypoxia in C. albicans is complex and is signaled both via lowered sterol levels and other unstudied mechanisms. We also show that induction of filamentation under hypoxic conditions requires the Ras1- and Cdc35-dependent pathway.

Combined Transfer Learning and Test-Time Augmentation Improves Convolutional Neural Network-Based Semantic Segmentation of Prostate Cancer from Multi-Parametric MR Images.

PURPOSE: Multiparametric MRI (mp-MRI) is a widely used tool for diagnosing and staging prostate cancer. The purpose of this study was to evaluate whether transfer learning, unsupervised pre-training and test-time augmentation significantly improved the performance of a convolutional neural network (CNN) for pixel-by-pixel prediction of cancer vs. non-cancer using mp-MRI datasets. METHODS: 154 subjects undergoing mp-MRI were prospectively recruited, 16 of whom subsequently underwent radical prostatectomy. Logistic regression, random forest and CNN models were trained on mp-MRI data using histopathology as the gold standard. Transfer learning, unsupervised pre-training and test-time augmentation were used to boost CNN performance. Models were evaluated using Dice score and area under the receiver operating curve (AUROC) with leave-one-subject-out cross validation. Permutation feature importance testing was performed to evaluate the relative value of each MR contrast to CNN model performance. Statistical significance (p<0.05) was determined using the paired Wilcoxon signed rank test with Benjamini-Hochberg correction for multiple comparisons. RESULTS: Baseline CNN outperformed logistic regression and random forest models. Transfer learning and unsupervised pre-training did not significantly improve CNN performance over baseline; however, test-time augmentation resulted in significantly higher Dice scores over both baseline CNN and CNN plus either of transfer learning or unsupervised pre-training. The best performing model was CNN with transfer learning and test-time augmentation (Dice score of 0.59 and AUROC of 0.93). The most important contrast was apparent diffusion coefficient (ADC), followed by Ktrans and T2, although each contributed significantly to classifier performance. CONCLUSIONS: The addition of transfer learning and test-time augmentation resulted in significant improvement in CNN segmentation performance in a small set of prostate cancer mp-MRI data. Results suggest that these techniques may be more broadly useful for the optimization of deep learning algorithms applied to the problem of semantic segmentation in biomedical image datasets. However, further work is needed to improve the generalizability of the specific model presented herein.

Prospective Validation of the Italian Alliance Against Cancer Lung Panel in Patients With Advanced Non-Small-Cell Lung Cancer.

BACKGROUND: The deeper knowledge of non-small-cell lung cancer (NSCLC) biology and the discovery of driver molecular alterations have opened the era of precision medicine in lung oncology, thus significantly revolutionizing the diagnostic and therapeutic approach to NSCLC. In Italy, however, molecular assessment remains heterogeneous across the country, and numbers of patients accessing personalized treatments remain relatively low. Nationwide programs have demonstrated that the creation of consortia represent a successful strategy to increase the number of patients with a molecular classification. PATIENTS AND METHODS: The Alliance Against Cancer (ACC), a network of 25 Italian Research Institutes, has developed a targeted sequencing panel for the detection of genomic alterations in 182 genes in patients with a diagnosis of NSCLC (ACC lung panel). One thousand metastatic NSCLC patients will be enrolled onto a prospective trial designed to measure the sensitivity and specificity of the ACC lung panel as a tool for molecular screening compared to standard methods. RESULTS AND CONCLUSION: The ongoing trial is part of a nationwide strategy of ACC to develop infrastructures and improve competences to make the Italian research institutes independent for genomic profiling of cancer patients.

Correlation between biofilm formation and the hypoxic response in Candida parapsilosis.

The ability of Candida parapsilosis to form biofilms on indwelling medical devices is correlated with virulence. To identify genes that are important for biofilm formation, we used arrays representing approximately 4,000 open reading frames (ORFs) to compare the transcriptional profile of biofilm cells growing in a microfermentor under continuous flow conditions with that of cells in planktonic culture. The expression of genes involved in fatty acid and ergosterol metabolism and in glycolysis, is upregulated in biofilms. The transcriptional profile of C. parapsilosis biofilm cells resembles that of Candida albicans cells grown under hypoxic conditions. We therefore subsequently used whole-genome arrays (representing 5,900 ORFs) to determine the hypoxic response of C. parapsilosis and showed that the levels of expression of genes involved in the ergosterol and glycolytic pathways, together with several cell wall genes, are increased. Our results indicate that there is substantial overlap between the hypoxic responses of C. parapsilosis and C. albicans and that this may be important for biofilm development. Knocking out an ortholog of the cell wall gene RBT1, whose expression is induced both in biofilms and under conditions of hypoxia in C. parapsilosis, reduces biofilm development.

The development of working memory spatialization revealed by using the cave paradigm in a two-alternative spatial choice.

When Western participants are asked to keep in mind a sequence of verbal items, they tend to associate the first items to the left and the last items to the right. This phenomenon, known as the spatial-positional association response codes effect, has been interpreted as showing that individuals spatialize the memoranda by creating a left-to-right mental line with them. One important gap in our knowledge concerns the development of this phenomenon: when do Western individuals start organizing their thought from left to right? To answer this question, 274 participants in seven age groups were tested (kindergarten, Grades 1, 2, 3, 4, and 5, and adults). We used a new protocol meant to be child-friendly, which involves associating two caves with two animals using a two-alternative spatial forced choice. Participants had to guess in which cave a specific animal could be hidden. Results showed that it is from Grade 3 on that participants spatialize information in working memory in a left-to-right fashion like adults.