Machine learning with the hierarchy-of-hypotheses (HoH) approach discovers novel pattern in studies on biological invasions

A Framework for Investigating Rules of Life by Establishing Zones of Influence

The incredible complexity of biological processes across temporal and spatial scales hampers defining common underlying mechanisms driving the patterns of life. However, recent advances in sequencing, big data analysis, machine learning, and molecular dynamics simulation have renewed the hope and urgency of finding potential hidden rules of life. There currently exists no framework to develop such synoptic investigations. Some efforts aim to identify unifying rules of life across hierarchical levels of time, space, and biological organization, but not all phenomena occur across all the levels of these hierarchies. Instead of identifying the same parameters and rules across levels, we posit that each level of a temporal and spatial scale and each level of biological organization has unique parameters and rules that may or may not predict outcomes in neighboring levels. We define this neighborhood, or the set of levels, across which a rule functions as the zone of influence. Here, we introduce the zone of influence framework and explain using three examples: (a) randomness in biology, where we use a Poisson process to describe processes from protein dynamics to DNA mutations to gene expressions, (b) island biogeography, and (c) animal coloration. The zone of influence framework may enable researchers to identify which levels are worth investigating for a particular phenomenon and reframe the narrative of searching for a unifying rule of life to the investigation of how, when, and where various rules of life operate.

Effect of Occupational Health and Safety Training for Chinese Construction Workers Based on the CHAID Decision Tree

Background: Occupational health and safety (OHS) training is an important way to prevent construction safety risks. However, the effectiveness of OHS training in China is questionable. In this study, the CHAID (chi-squared automatic interaction detection) decision tree, chi-square analysis, and correlation analysis were used to explore the main, secondary, weak, unrelated, and expectation factors affecting the effectiveness of training. It is the first to put forward the "five-factor method" of training effectiveness. It is found that training effectiveness is positively correlated with job responsibilities, OHS training, and job satisfaction. It is also significantly related to job certificate, training time, training method, and working time. However, the effectiveness of training has nothing to do with personal age, marital status, educational level, job type, and whether or not they have experienced industrial accidents. And the workers on site expect the enterprise to provide security and opportunities such as physical safety, training and learning, and future career development. The results show that OHS system training should be strengthened in the construction industry, and classified training should be carried out according to post responsibility, training methods, job satisfaction, and working hours.

The Hierarchy-of-Hypotheses Approach: A Synthesis Method for Enhancing Theory Development in Ecology and Evolution

In the current era of Big Data, existing synthesis tools such as formal meta-analyses are critical means to handle the deluge of information. However, there is a need for complementary tools that help to (a) organize evidence, (b) organize theory, and (c) closely connect evidence to theory. We present the hierarchy-of-hypotheses (HoH) approach to address these issues. In an HoH, hypotheses are conceptually and visually structured in a hierarchically nested way where the lower branches can be directly connected to empirical results. Used for organizing evidence, this tool allows researchers to conceptually connect empirical results derived through diverse approaches and to reveal under which circumstances hypotheses are applicable. Used for organizing theory, it allows researchers to uncover mechanistic components of hypotheses and previously neglected conceptual connections. In the present article, we offer guidance on how to build an HoH, provide examples from population and evolutionary biology and propose terminological clarifications.

Machine learning with the hierarchy-of-hypotheses (HoH) approach discovers novel pattern in studies on biological invasions

Research synthesis on simple yet general hypotheses and ideas is challenging in scientific disciplines studying highly context-dependent systems such as medical, social, and biological sciences. This study shows that machine learning, equation-free statistical modeling of artificial intelligence, is a promising synthesis tool for discovering novel patterns and the source of controversy in a general hypothesis. We apply a decision tree algorithm, assuming that evidence from various contexts can be adequately integrated in a hierarchically nested structure. As a case study, we analyzed 163 articles that studied a prominent hypothesis in invasion biology, the enemy release hypothesis. We explored if any of the nine attributes that classify each study can differentiate conclusions as classification problem. Results corroborated that machine learning can be useful for research synthesis, as the algorithm could detect patterns that had been already focused in previous narrative reviews. Compared with the previous synthesis study that assessed the same evidence collection based on experts' judgement, the algorithm has newly proposed that the studies focusing on Asian regions mostly supported the hypothesis, suggesting that more detailed investigations in these regions can enhance our understanding of the hypothesis. We suggest that machine learning algorithms can be a promising synthesis tool especially where studies (a) reformulate a general hypothesis from different perspectives, (b) use different methods or variables, or (c) report insufficient information for conducting meta-analyses.

Do cancer stem cells exist? A pilot study combining a systematic review with the hierarchy-of-hypotheses approach

The phenomenon of cancer cell heterogeneity has been explained by different hypotheses, each entailing different therapy strategies. The most recent is the cancer stem cell model, which says that tumourigenicity and self-renewal are restricted to rare stem cell-like cancer cells. Since its conception, conflicting evidence has been published. In this study, we tested the applicability of a new approach developed in the field of ecology, the hierarchy-of-hypotheses approach, for the Cancer Stem Cell hypothesis. This approach allows to structure a broad concept into more specific sub-hypotheses, which in turn can be connected to available empirical studies. To generate a dataset with empirical studies, we conducted a systematic literature review in the Web of Science limited to the first 1000 publications returned by the search. From this pool, 51 publications were identified that tested whether a cell sub-population had cancer stem cell properties. By classifying the studies according to: (1) assessed indicators, (2) experimental assays and (3) model cancer cells used, we built a hierarchical structure of sub-hypotheses. The empirical tests from the selected studies were subsequently assigned to this hierarchy of hypotheses, and the percentage of supporting, undecided and questioning evidence was calculated for each sub-hypothesis, as well as additional experimental characteristics. Our approach successfully allowed us to determine that within our dataset, the empirical support for the CSC hypothesis was only 49.0%. The support of different sub-hypotheses was highly variable. Most noticeable, the conception that putative cancer stem cells are a rare subset of cells could not be confirmed by most studies (13.5% support). The empirical support varied also between types of cancer, animal models and cell isolation method used. For the first time, this study showed the applicability of the hierarchy-of-hypotheses approach for synthesizing and evaluating empirical evidence for a broad hypothesis in the field of bio-medical research.