Understanding the complex mechanisms of drugs' therapeutic effects is essential for advancing precision medicine and optimizing treatment strategies. However, systematically distinguishing between drugs that address root causes (etiological mechanisms) and those that alleviate symptoms (palliative mechanisms) using modern Artificial Intelligence (AI)-based strategies remains underexplored. We present a novel computational framework for classifying drug Mechanisms of Action (MoA) as etiological or palliative, comparing three approaches: (i) Fine-tuning Science Bidirectional Encoder Representations from Transformers (SciBERT) with drug descriptions; (ii) Training various Graph Neural Networks (GNNs) on a constructed heterogeneous network of drugs, genes, and diseases; and (iii) Developing DruGNNosis-MoA, which integrates GNN with our fine-tuned SciBERT embeddings as node features. DruGNNosis-MoA excelled (F1-score 0.94) at identifying drug MoA. DruGNNosis-MoA characterizes drug mechanisms for subsequent pharmacological studies, thereby advancing precision medicine and therapeutic development.

MOTIVATION: Micro-blogging with Twitter to communicate new results, discuss ideas and share techniques is becoming central. While most Twitter users are real people, the Twitter API provides the opportunity to develop Twitter bots and to analyze global trends in tweets. RESULTS: EnrichrBot is a bot that tracks and tweets information about human genes implementing six principal functions: (i) tweeting information about under-studied genes including non-coding lncRNAs, (ii) replying to requests for information about genes, (iii) responding to GWASbot, another bot that tweets Manhattan plots from genome-wide association study analysis of the UK Biobank, (iv) tweeting randomly selected gene sets from the Enrichr database for analysis with Enrichr, (v) responding to mentions of human genes in tweets with additional information about these genes and (vi) tweeting a weekly report about the most trending genes on Twitter. AVAILABILITY AND IMPLEMENTATION: https://twitter.com/botenrichr; source code: https://github.com/MaayanLab/EnrichrBot. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

People's social life has become more embedded in dynamic online communities. Each online community can be viewed as a temporal online social network (OSN). The interaction level among OSN members leads to the emergence of dynamic social roles, which change and evolve over time, creating a sequence of temporal roles. These role sequences show diversity in the role-affiliation frequency of members. That diversity enables modeling the dynamic behaviors of individuals. This paper proposes a temporal role-affiliation frequency model (RAFM) which detects the time evolving roles of each member and analyzes her/his role-affiliation frequency to infer her/his latent behavior. Applying the RAFM to real interaction data, collected in four online communities, revealed the identity of influential members. In addition, members with similar temporal behavioral patterns were found to have similar latent behavior patterns. These patterns are manifested via similar role transitions in different OSNs whose temporal interaction rhythms were compatible. These two research findings contribute to OSN research and knowledge via improved understanding of member behavior online based on role-affiliation frequency and role transitions. Thus, member latent behavior can be inferred, and influential members can be identified.

Exposure to online participation in an online social network (OSN) is often modeled as a neighbor-to-neighbor flow. Exposure to participation shift of a neighbor in an OSN, can influence a user to shift participation and engage in posting. Since many postings on OSNs are public, exposure to participation shifts of non-neighbors can also influence one's posting engagement. Modeling local and global influence on the posting engagement of a user, we demonstrate the model on Twitter interactions related to Gaza's Great Return March (GRM) and show the effects of both local and global influence.

Lately there is great progress in business organizations perception towards social aspects. Competitive organizations need to create innovation and segregate in the market. Business interactions help reaching those goals but finding the effective interactions is a challenge. We propose a prediction method, based on social networks analysis (SNA) and text data mining (TDM), for predicting which nodes in a social network will be linked next. The network which is used to demonstrate the proposed prediction method is composed of academic co-authors who collaborated on writing articles. Without loss of generality, the academic co-authoring network demonstrates the proposed prediction procedure due to its similarity to other networks, such as business co-operation networks. The results show that the best prediction is achieved by incorporating TDM with SNA.