Over the last several decades, researchers have time and again proposed use of non-chromatographic methods for processing of biotherapeutic products. However, chromatography continues to be the backbone of downstream processing, particularly at process scale. There are many reasons for this, critical ones being the unparalleled scalability, robustness, and selectivity that process chromatography offers over its peers. It is no surprise then that process chromatography has been a topic of major developments in resin matrix, ligand chemistry, modalities, high throughput process development, process modelling, and approaches for control. In this review, we attempt to summarize major developments in the above-mentioned areas. Greater significance has been given to advancements in the last 5 years (2013–2017).

Biopharmaceutical industry is presently facing multiple economic, political and regulatory challenges. The adoption of continuous bioprocessing is expected to partly alleviate these. Over the years, many technology solutions have been developed for various biotech unit operations which were hitherto operated in batch fashion. The next challenge is to integrate all unit operations to create a fully continuous manufacturing system. In this article, we discuss recent developments in process integration and control under a proposed framework consisting of modular, adaptation and merger approaches of integration with a particular emphasis on developments that have occurred in the past five years (2013-2018).

A correction to this article has been published and is linked from the HTML and PDF versions of this paper. The error has not been fixed in the paper.

Thermoelectric power production in the United States primarily relies on wet-cooled plants, which in turn require water below prescribed design temperatures, both for cooling and operational efficiency. Thus, power production in US remains particularly vulnerable to water scarcity and rising stream temperatures under climate change and variability. Previous studies on the climate-water-energy nexus have primarily focused on mid- to end-century horizons and have not considered the full range of uncertainty in climate projections. Technology managers and energy policy makers are increasingly interested in the decadal time scales to understand adaptation challenges and investment strategies. Here we develop a new approach that relies on a novel multivariate water stress index, which considers the joint probability of warmer and scarcer water, and computes uncertainties arising from climate model imperfections and intrinsic variability. Our assessments over contiguous US suggest consistent increase in water stress for power production with about 27% of the production severely impacted by 2030s.

Characterization of climate uncertainty at regional scales over near-term planning horizons (0–30 years) is crucial for climate adaptation. Climate internal variability (CIV) dominates climate uncertainty over decadal prediction horizons at stakeholders’ scales (regional to local). In the literature, CIV has been characterized indirectly using projections of climate change from multi-model ensembles (MME) instead of directly using projections from multiple initial condition ensembles (MICE), primarily because adequate number of initial condition (IC) runs were not available for any climate model. Nevertheless, the recent availability of significant number of IC runs from one climate model allows for the first time to characterize CIV directly from climate model projections and perform a sensitivity analysis to study the dominance of CIV compared to model response variability (MRV). Here, we measure relative agreement (a dimensionless number with values ranging between 0 and 1, inclusive; a high value indicates less variability and vice versa) among MME and MICE and find that CIV is lower than MRV for all projection time horizons and spatial resolutions for precipitation and temperature. However, CIV exhibits greater dominance over MRV for seasonal and annual mean precipitation at higher latitudes where signals of climate change are expected to emerge sooner. Furthermore, precipitation exhibits large uncertainties and a rapid decline in relative agreement from global to continental, regional, or local scales for MICE compared to MME. The fractional contribution of uncertainty due to CIV is invariant for precipitation and decreases for temperature as lead time progresses towards the end of the century.

Achieving consistent product quality of a biotherapeutic is a major target for any biopharmaceutical manufacturer, even more for a biosimilar producer as comparability with the innovator product is a regulatory expectation. The complexity of biotherapeutic products and their tedious manufacturing processes, however, make this a non-trivial exercise. The primary motivation of this work is to develop an integrated chromatographic platform for purification of monoclonal antibody (mAb) therapeutics that can deliver the desired separation of both charge variants and aggregates, in addition to the process related impurities like host cell proteins (HCP) and host cell DNA. To achieve the same, an integrated two-stage chromatographic process platform consisting of cation exchange chromatography and multimodal chromatography is being proposed. The versatility of the proposed platform has been successfully demonstrated for three different mAbs. It have been shown that in each case charge variant separation is achieved with the required clearance of aggregates (<1%), HCP (<10 ppm), and DNA (<5 ppb). Moreover, the proposed platform is conducive to use for development of a continuous process and offers smaller process time, lower buffer utilization, and decreased operational costs when compared to the conventional purification platforms.

Current evaluations of GCMs, as detailed by the IPCC AR5, show that different GCMs often illustrate different skills for different tasks. The IPCC uses an average or multi-model ensemble (MME) for future projections. In this work, we consider evaluating and combining GCMs for multiple tasks from a multi-task learning (MTL) perspective. MTL is an approach to machine learning that learns multiple problems simultaneously thus leading to a better model by taking advantage of the commonality among the tasks [1], [2]. But unlike traditional MTL, for our problem the task relationships are not known initially. Further, each GCM has multiple runs corresponding to different initial conditions which need to be suitably considered. In this work, we present approaches to multitask sparse structure learning (MSSL) which estimate task relationships along with learning suitable multi-model combinations for each task, and explore ideas for handling multiple initial condition runs for GCMs.

Climate projections, especially at decadal to century scales, rely on physics-based computer models. While the models have generated useful information about global warming and hydrology, "the sad truth of climate science is that the most crucial information is the least credible" (Schiermeier 2010). The possible reasons include intrinsic variability of the climate system as well as our lack of understanding of the physics and the inability to include known physics within the current generation of computer models. Data sciences, ranging from statistics and signal processing to machine learning and nonlinear dynamics, continue to help fill some of the crucial gaps in climate. We hypothesize that these data science solutions can be improved if they are driven by physical knowledge, especially when this knowledge cannot be incorporated in current climate models because they are incomplete or incompatible. We have called this paradigm Physics-Guided Data Mining (PGDM) (Ganguly et al. 2014). In addition to motivating and introducing PGDM, this chapter presents three case studies on precipitation, based on our prior work. Statistical downscaling, which generates higher-resolution projections from lower-resolution model simulations, benefits from a blend of sparse learning techniques with physically motivated covariates (Das et al. 2014; 2015). Multimodal uncertainty quantification shows the potential to improve when physical relations with ancillary variables are considered together with historical skills and future consensus, within a Bayesian framework (Ganguly et al. 2013; Kodra 2014a; Smith et al. 2009). Characterization of internal variability and associated model performance benefit from data-driven analysis of multi-initial condition ensembles, combined with physical understanding of oceanic indices and their initializations (Kodra et al. 2012). The proposed PGDM paradigm, illustrated through our prior publications, shows the potential to bridge crucial knowledge gaps in climate science and help in translation to water resources impacts.

Climate models are thought to solve boundary value problems unlike numerical weather prediction, which is an initial value problem. However, climate internal variability (CIV) is thought to be relatively important at near-term (0-30 year) prediction horizons, especially at higher resolutions. The recent availability of significant numbers of multi-model (MME) and multi-initial condition (MICE) ensembles allows for the first time a direct sensitivity analysis of CIV versus model response variability (MRV). Understanding the relative agreement and variability of MME and MICE ensembles for multiple regions, resolutions, and projection horizons is critical for focusing model improvements, diagnostics, and prognosis, as well as impacts, adaptation, and vulnerability studies. Here we find that CIV (MICE agreement) is lower (higher) than MRV (MME agreement) across all spatial resolutions and projection time horizons for both temperature and precipitation. However, CIV dominates MRV over higher latitudes generally and in specific regions. Furthermore, CIV is considerably larger than MRV for precipitation compared to temperature across all horizontal and projection scales and seasons. Precipitation exhibits larger uncertainties, sharper decay of MICE agreement compared to MME, and relatively greater dominance of CIV over MRV at higher latitudes. The findings are crucial for climate predictability and adaptation strategies at stakeholder-relevant scales.

The structure, interdependence, and fragility of systems ranging from power-grids and transportation to ecology, climate, biology and even human communities and the Internet have been examined through network science. While response to perturbations has been quantified, recovery strategies for perturbed networks have usually been either discussed conceptually or through anecdotal case studies. Here we develop a network science based quantitative framework for measuring, comparing and interpreting hazard responses as well as recovery strategies. The framework, motivated by the recently proposed temporal resilience paradigm, is demonstrated with the Indian Railways Network. Simulations inspired by the 2004 Indian Ocean Tsunami and the 2012 North Indian blackout as well as a cyber-physical attack scenario illustrate hazard responses and effectiveness of proposed recovery strategies. Multiple metrics are used to generate various recovery strategies, which are simply sequences in which system components should be recovered after a disruption. Quantitative evaluation of these strategies suggests that faster and more efficient recovery is possible through network centrality measures. Optimal recovery strategies may be different per hazard, per community within a network, and for different measures of partial recovery. In addition, topological characterization provides a means for interpreting the comparative performance of proposed recovery strategies. The methods can be directly extended to other Large-Scale Critical Lifeline Infrastructure Networks including transportation, water, energy and communications systems that are threatened by natural or human-induced hazards, including cascading failures. Furthermore, the quantitative framework developed here can generalize across natural, engineered and human systems, offering an actionable and generalizable approach for emergency management in particular as well as for network resilience in general.

Resilience to nonstationarity and deep uncertainty is a prerequisite to water security. Stakeholder planning horizons typically extend to about 30 years in water quantity or quality management, flood or drought hazard resilience, or the water-energy-food-ecosystems nexus. Projections of stressors, such as population, land use, stability assumptions of technologies, infrastructures, and organizations, are relatively more credible at the nearer term. However, compared to longer lead times of mid- to end-century and beyond, climate adaptation challenges are more acute. Over 30-year horizons, the degree of nonstationarity is comparable to the overall uncertainty, which in turn is dominated by natural variability, especially at higher space-time resolutions. A case study with power production at risk in the US suggests that informed decisions could be possible despite nonstationarity and deep uncertainty.

Abstract Spatial and spatiotemporal data mining is the process of discovering interesting and previously unknown, but potentially useful patterns from the data collected over time and space. However, explosive growth in the spatial and spatiotemporal data, and the emergence of social media and location sensing technologies, emphasizes the need for developing new and computationally efficient methods tailored for analyzing big data. In this chapter, we study approaches to handle big spatial and spatiotemporal data by closely looking at the computational and I/O requirements of several analysis algorithms for such data. We also study applications of such methods in domains where data is encountered at a massive scale.

Wind extremes have consequences for renewable energy sectors, critical infrastructures, coastal ecosystems, and insurance industry. Considerable debates remain regarding the impacts of climate change on wind extremes. While climate models have occasionally shown increases in regional wind extremes, a decline in the magnitude of mean and extreme near-surface wind speeds has been recently reported over most regions of the Northern Hemisphere using observed data. Previous studies of wind extremes under climate change have focused on selected regions and employed outputs from the regional climate models (RCMs). However, RCMs ultimately rely on the outputs of global circulation models (GCMs), and the value-addition from the former over the latter has been questioned. Regional model runs rarely employ the full suite of GCM ensembles, and hence may not be able to encapsulate the most likely projections or their variability. Here we evaluate the performance of the latest generation of GCMs, the Coupled Model Intercomparison Project phase 5 (CMIP5), in simulating extreme winds. We find that the multimodel ensemble (MME) mean captures the spatial variability of annual maximum wind speeds over most regions except over the mountainous terrains. However, the historical temporal trends in annual maximum wind speeds for the reanalysis data, ERA-Interim, are not well represented in the GCMs. The historical trends in extreme winds from GCMs are statistically not significant over most regions. The MME model simulates the spatial patterns of extreme winds for 25–100 year return periods. The projected extreme winds from GCMs exhibit statistically less significant trends compared to the historical reference period.

Spatial and spatiotemporal data mining is the process of discovering interesting and previously unknown, but potentially useful patterns from the data collected over time and space. However, explosive growth in the spatial and spatiotemporal data, and the emergence of social media and location sensing technologies, emphasizes the need for developing new and computationally efficient methods tailored for analyzing big data. In this chapter, we study approaches to handle big spatial and spatiotemporal data by closely looking at the computational and I/O requirements of several analysis algorithms for such data. We also study applications of such methods in domains where data is encountered at a massive scale.

Regional and seasonal temperature and precipitation over land are compared across two generations of global climate model ensembles, specifically, CMIP5 and CMIP3, through historical twentieth century skills and multi-model agreement, and twenty first century projections. A suite of diagnostic and performance metrics, ranging from spatial bias or model-consensus maps and aggregate time series plots, to measures of equivalence between probability density functions and Taylor diagrams, are used for the intercomparisons. Pairwise and multi-model ensemble comparisons were performed for 11 models, which were selected based on data availability and resolutions. Results suggest little change in the central tendency or variability or uncertainty of historical skills or consensus across the two generations of models. However, there are regions and seasons, at different levels of aggregation, where significant changes, performance improvements, and even degradation in skills, are suggested. The insights may provide directions for further improvements in next generations of climate models, and in the meantime, help inform adaptation and policy.