Efficient pre-processing of Single-cell ATAC-seq data

The primary tool currently used to pre-process 10X Chromium single-cell ATAC-seq data is Cell Ranger, which can take very long to run on standard datasets. To facilitate rapid pre-processing that enables reproducible workflows, we present a suite of tools called scATAK for pre-processing single-cell ATAC-seq data that is 18 times faster than Cell Ranger on human samples, and that uses 33% less RAM when 8 CPU threads are used. Our tool can also calculate chromatin interaction potential matrices, and generate open chromatin signals and interaction traces for cell groups. We demonstrate the utility of scATAK in an exploration of the chromatin regulatory landscape of a healthy adult human brain and show that it can reveal cell-type-specific features. scATAK is available at https://pachterlab.github.io/scATAK/.

Theorems for free from separation logic specifications

Separation logic specifications with abstract predicates intuitively enforce a discipline that constrains when and how calls may be made between a client and a library. Thus a separation logic specification of a library intuitively enforces a protocol on the trace of interactions between a client and the library. We show how to formalize this intuition and demonstrate how to derive "free theorems" about such interaction traces from abstract separation logic specifications. We present several examples of free theorems. In particular, we prove that a so-called logically atomic concurrent separation logic specification of a concurrent module operation implies that the operation is linearizable. All the results presented in this paper have been mechanized and formally proved in the Coq proof assistant using the Iris higher-order concurrent separation logic framework.

Adaptive Help System Based on Learners ‘Digital Traces’ and Learning Styles

Learning management system (LMS) such as Claroline, Ganesha, Chamilo, Moodle ..., are commonly and well used in e-education (e-learning). Most of theTechnology Enhanced Learning (TEL) focus on supporting teachers in the creation and organization of online courses. However, in general, they do not consider individual differences of each learner. In addition, they do not provide enough indicators which will help to track the learners. In this paper, we investigate the benefits of integrating learning styles in the Web-based educational systems. Also we are interested in the use of interaction traces in order to address the lack of feedback between the learner and the teacher. Generally, we aim to offer a tool that allows the tutor and the instructional designer to interpret learner courses, in order to provide help as needed for each individual.

Impact of GUI personalization of a word processor on a learning activity course

This research investigates the impact of an activity of personalization of a graphical user interface by the learners, on their behavior of using the ILE. The analysis conducted is based on an exploitation of the interaction traces between the learner and the interface of a word processor software with advanced personalization and auto-writing features including training of spelling and a learning analytics management module. The results show that, several variables related to the facilitation conditions recognized by the ILE partly explain the writing activity. Navigation variable can be correlated with the knowledge of customization possibilities. If the automatic sentence generator has no significant effect on the number of misspellings found in the documents submitted, the intention to personalize the interface seems to have a greater effect than the act of personalization itself. But the impact of the personalization process on learning outcomes is still to be established.