VC@Scale: Scalable and high-performance variant calling on cluster environments.

BACKGROUND: Recently many new deep learning-based variant-calling methods like DeepVariant have emerged as more accurate compared with conventional variant-calling algorithms such as GATK HaplotypeCaller, Sterlka2, and Freebayes albeit at higher computational costs. Therefore, there is a need for more scalable and higher performance workflows of these deep learning methods. Almost all existing cluster-scaled variant-calling workflows that use Apache Spark/Hadoop as big data frameworks loosely integrate existing single-node pre-processing and variant-calling applications. Using Apache Spark just for distributing/scheduling data among loosely coupled applications or using I/O-based storage for storing the output of intermediate applications does not exploit the full benefit of Apache Spark in-memory processing. To achieve this, we propose a native Spark-based workflow that uses Python and Apache Arrow to enable efficient transfer of data between different workflow stages. This benefits from the ease of programmability of Python and the high efficiency of Arrow's columnar in-memory data transformations. RESULTS: Here we present a scalable, parallel, and efficient implementation of next-generation sequencing data pre-processing and variant-calling workflows. Our design tightly integrates most pre-processing workflow stages, using Spark built-in functions to sort reads by coordinates and mark duplicates efficiently. Our approach outperforms state-of-the-art implementations by >2 times for the pre-processing stages, creating a scalable and high-performance solution for DeepVariant for both CPU-only and CPU + GPU clusters. CONCLUSIONS: We show the feasibility and easy scalability of our approach to achieve high performance and efficient resource utilization for variant-calling analysis on high-performance computing clusters using the standardized Apache Arrow data representations. All codes, scripts, and configurations used to run our implementations are publicly available and open sourced; see https://github.com/abs-tudelft/variant-calling-at-scale.

Optimizing performance of GATK workflows using Apache Arrow In-Memory data framework.

BACKGROUND: Immense improvements in sequencing technologies enable producing large amounts of high throughput and cost effective next-generation sequencing (NGS) data. This data needs to be processed efficiently for further downstream analyses. Computing systems need this large amounts of data closer to the processor (with low latency) for fast and efficient processing. However, existing workflows depend heavily on disk storage and access, to process this data incurs huge disk I/O overheads. Previously, due to the cost, volatility and other physical constraints of DRAM memory, it was not feasible to place large amounts of working data sets in memory. However, recent developments in storage-class memory and non-volatile memory technologies have enabled computing systems to place huge data in memory to process it directly from memory to avoid disk I/O bottlenecks. To exploit the benefits of such memory systems efficiently, proper formatted data placement in memory and its high throughput access is necessary by avoiding (de)-serialization and copy overheads in between processes. For this purpose, we use the newly developed Apache Arrow, a cross-language development framework that provides language-independent columnar in-memory data format for efficient in-memory big data analytics. This allows genomics applications developed in different programming languages to communicate in-memory without having to access disk storage and avoiding (de)-serialization and copy overheads. IMPLEMENTATION: We integrate Apache Arrow in-memory based Sequence Alignment/Map (SAM) format and its shared memory objects store library in widely used genomics high throughput data processing applications like BWA-MEM, Picard and GATK to allow in-memory communication between these applications. In addition, this also allows us to exploit the cache locality of tabular data and parallel processing capabilities through shared memory objects. RESULTS: Our implementation shows that adopting in-memory SAM representation in genomics high throughput data processing applications results in better system resource utilization, low number of memory accesses due to high cache locality exploitation and parallel scalability due to shared memory objects. Our implementation focuses on the GATK best practices recommended workflows for germline analysis on whole genome sequencing (WGS) and whole exome sequencing (WES) data sets. We compare a number of existing in-memory data placing and sharing techniques like ramDisk and Unix pipes to show how columnar in-memory data representation outperforms both. We achieve a speedup of 4.85x and 4.76x for WGS and WES data, respectively, in overall execution time of variant calling workflows. Similarly, a speedup of 1.45x and 1.27x for these data sets, respectively, is achieved, as compared to the second fastest workflow. In some individual tools, particularly in sorting, duplicates removal and base quality score recalibration the speedup is even more promising. AVAILABILITY: The code and scripts used in our experiments are available in both container and repository form at: https://github.com/abs-tudelft/ArrowSAM .

A rare case of Waterhouse- Friderichsen syndrome during primary Varicella zoster infection.

Primary Varicella zoster virus infection in adults is associated with a higher risk of complications when compared with the benign disease course of primary infection during childhood. We present a rare complication of adult primary Varicella zoster in the form of acute, irreversible adrenal insufficiency due to bilateral adrenal haemorrhage, which is also known as the WaterhouseFriderichsensyndrome.

No added value of the age-adjusted D-dimer cut-off to the YEARS algorithm in patients with suspected pulmonary embolism.

Essentials Imaging is warranted in the majority of patients to confirm or rule out pulmonary embolism (PE). The age-adjusted D-dimer (ADJUST) reduced the number of required imaging tests in patients ≥ 50 years. The YEARS algorithm was designed to improve the efficiency in patients with suspected PE. There was no added value of implementing ADJUST in the YEARS algorithm in our cohort. SUMMARY: Background The YEARS algorithm was designed to simplify the diagnostic work-up of pulmonary embolism (PE) and to reduce the number of necessary computed tomography pulmonary angiography (CTPA) scans. An alternative strategy to reduce the number of CTPAs is the age-adjusted D-dimer cut-off (ADJUST) in patients aged 50 years or older. We aimed to investigate whether a combination of both diagnostic strategies might save additional CTPAs. Methods The YEARS algorithm consists of three items (clinical signs of deep venous thrombosis, hemoptysis, 'PE most likely diagnosis') with simultaneous D-dimer testing using a pre-test dependent threshold. We performed a post hoc analysis in 3465 patients managed according to YEARS to compare the number of patients managed without CTPA scans and associated diagnostic failures in hypothetical scenarios with different YEARS-ADJUST combinations. Results Following the YEARS algorithm, 1651 patients (48%) were managed without CTPA; PE was diagnosed in 456 (13%) patients at baseline and 18 patients with initial normal testing suffered venous thromboembolism (VTE) during 3-month follow-up (failure rate 0.61%; 95% confidence interval [CI], 0.36-0.96). If ADJUST had been fully integrated in YEARS, 1627 patients (47%) would have been managed without CTPA (absolute decrease of 0.69%; 95% CI -1.7 to 3.0), at cost of four additional missed PE diagnoses at baseline, for a projected 3-month VTE failure rate of 0.75% (95% CI, 0.49-1.13). None of the other studied scenarios showed relevant improvements in efficiency as well, but all led to more missed diagnoses. Conclusion In our cohort, there was no added value of implementing ADJUST in the YEARS algorithm.

Nailfold capillary abnormalities in sclerodermatous chronic GVHD.

Chronic GVHD (cGVHD) complicating allo-SCT commonly presents as sclerotic skin changes resembling systemic sclerosis (SSc), suggesting a common pathophysiological pathway. Damage to capillaries is considered an early event in the pathogenesis of SSc, and is associated with characteristic nailfold capillary abnormalities. Whether such nailfold capillary abnormalities occur in sclerodermatous cGVHD is unknown. Nailfold videocapillaroscopy (NVC) was used to evaluate capillary morphology, density and loop dimensions in 14 patients with sclerodermatous cGVHD, 14 sex- and age-matched SSc patients, and 14 healthy controls. It was shown that none of the cGVHD patients and controls, whereas all SSc patients showed severe capillary abnormalities. cGVHD patients and controls showed no differences in capillary density (9.05 vs 9.16 loops/mm, respectively, P=0.84), and capillary loop dimensions (total loop width 44.36 vs 45.56 µm, respectively, P=0.84). Compared with cGVHD patients, SSc patients had a reduced capillary density (9.05 vs 5.25 loops/mm, respectively, P<0.001), and an increase in capillary loop dimensions (total loop width 44.36 vs 99.97 µm, respectively, P=<0.001). In conclusion sclerodermatous cGVHD patients do not show the characteristic microvascular abnormalities seen in SSc, suggesting that capillary damage does not contribute to the pathophysiology of sclerodermatous cGVHD, and making NVC unsuitable for early identification.

Hestia criteria can safely select patients with pulmonary embolism for outpatient treatment irrespective of right ventricular function.

BACKGROUND: There has been debate over how patients with pulmonary embolism (PE) can be safely selected for outpatient treatment. OBJECTIVES: To compare the Hestia criteria with the European Society of Cardiology (ESC) criteria for selecting low-risk patients with PE for outpatient treatment. METHODS: From 2008 to 2010, 496 patients with acute, symptomatic PE were screened and 275 treated at home and 221 treated in the hospital according to the Hestia Study protocol. The Hestia criteria were used to select patients for outpatient treatment. Right and left ventricular (RV and LV) diameters were measured on computed tomography images. RV dysfunction was defined as an RV/LV ratio > 1.0. Patients were classified according to the ESC criteria into low, intermediate and high-risk groups, based on blood pressure and RV dysfunction. During 3 months follow-up adverse events were scored. RESULTS: Adverse events occurred in 22 patients (4.5%) treated in the hospital vs. none of the patients treated at home (P < 0.001). Sensitivity and negative predictive value for adverse outcome were 100% for the Hestia criteria and 96% and 99% for the ESC criteria, respectively. Of the patients treated at home according to the Hestia criteria, 35% were normotensive but had RV dysfunction and were classified as intermediate risk according to the ESC criteria. No adverse events happened in these patients treated at home. CONCLUSIONS: Clinical criteria, such as the Hestia criteria, could be helpful in selecting patients, including those with RV dysfunction who have a low risk of adverse clinical outcome and could be candidates for outpatient treatment.

Outpatient treatment in patients with acute pulmonary embolism: the Hestia Study.

BACKGROUND: Traditionally, patients with pulmonary embolism (PE) are initially treated in the hospital with low molecular weight heparin (LMWH). The results of a few small non-randomized studies suggest that, in selected patients with proven PE, outpatient treatment is potentially feasible and safe. OBJECTIVE: To evaluate the efficacy and safety of outpatient treatment according to predefined criteria in patients with acute PE. PATIENTS AND METHODS: A prospective cohort study of patients with objectively proven acute PE was conducted in 12 hospitals in The Netherlands between 2008 and 2010. Patients with acute PE were triaged with the predefined criteria for eligibility for outpatient treatment, with LMWH (nadroparin) followed by vitamin K antagonists. All patients eligible for outpatient treatment were sent home either immediately or within 24 h after PE was objectively diagnosed. Outpatient treatment was evaluated with respect to recurrent venous thromboembolism (VTE), including PE or deep vein thrombosis (DVT), major hemorrhage and total mortality during 3 months of follow-up. RESULTS: Of 297 included patients, who all completed the follow-up, six (2.0%; 95% confidence interval [CI] 0.8-4.3) had recurrent VTE (five PE [1.7%] and one DVT [0.3%]). Three patients (1.0%, 95% CI 0.2-2.9) died during the 3 months of follow-up, none of fatal PE. Two patients had a major bleeding event, one of which was fatal intracranial bleeding (0.7%, 95% CI 0.08-2.4). CONCLUSION: Patients with PE selected for outpatient treatment with predefined criteria can be treated with anticoagulants on an outpatient basis. (Dutch Trial Register No 1319; http://www.trialregister.nl/trialreg/index.asp).

Nailfold capillary density is associated with the presence and severity of pulmonary arterial hypertension in systemic sclerosis.

OBJECTIVE: The aim of this study was to investigate whether there are differences in capillary nailfold changes in patients with systemic sclerosis (SSc) with and without pulmonary arterial hypertension (PAH), and whether these changes are associated with PAH severity and disease specificity. METHODS: Capillary density and loop dimensions were studied in 21 healthy controls, 20 patients with idiopathic PAH (IPAH) and 40 patients with SSc. Of the 40 patients with SSc, 19 had no PAH (SSc-nonPAH) and 21 had PAH (SSc-PAH), of whom eight had PAH during exercise. RESULTS: Capillary density was lower in SSc-PAH compared with patients who had SSc-nonPAH (4.33/mm vs 6.56/mm respectively, p = 0.001), but loop dimensions were equal. In comparison with IPAH, patients with SSc-PAH had reduced capillary density (4.33/mm vs 7.86/mm, p<0.001) and larger loop dimensions (total width 101.05 microm vs 44.43 microm, p<0.001). Capillary density in healthy controls (9.87/mm) was significantly higher when compared with SSc-nonPAH (6.56/mm), SSc-PAH (4.33/mm) and with IPAH (7.86/mm). No differences in capillary dimensions were present between healthy controls and IPAH. Capillary density correlated with mean pulmonary arterial pressure (PAP) at rest in SSc-PAH at rest (r = -0.58, p = 0.039) and IPAH (r = -0.67, p = 0.001). CONCLUSIONS: Reduction of nailfold capillary density, but not capillary loop dimensions is associated with PAH, and correlates with the severity of PAH in both SSc and IPAH. This suggests that either systemic microvascular changes play a part in the development of PAH, or that PAH itself contributes to systemic microvascular changes.

[Diagnostic image (70). A man with fever, swollen joints and erythema nodosum. Loefgren's syndrome]. / Diagnose in beeld (70). Een man met koorts, gewrichtszwelling en nodulair erytheem. Syndroom van Löfgren.

A 39-year-old man presented with fever, arthritis of knees and wrists, periarticular ankle inflammation, erythema nodosum, bilateral hilar adenopathy and diffuse pulmonary parenchymal changes, due to Löfgren's syndrome.

The role of BCL-2 and bax protein in monocyte-mediated apoptosis in human leukemic cell lines.

Monocytes or monocyte-derived supernatants are able to kill leukemic cells via apoptosis, thereby preferentially effecting more mature leukemic cells. In the present study, the relationship between apoptosis and the apoptosis related proteins, bcl-2 and bax, was investigated in a number of human leukemic cell lines. Monocyte-derived supernatant induces extensive apoptosis in U937 myeloid leukemia cells and minor apoptosis in HL60 cells. No apoptosis was seen in four other cell lines (THP1, HL60-D3, KG1, and K562). The expression of bcl-2 and bax protein was determined in both groups of leukemic cell lines by flow cytometry (bcl-2 and bax) and Western blotting (bcl-2) at baseline level and after incubation with monocyte supernatant after different time periods. No clear relation was found between baseline bcl-2 or bax protein expression and the occurrence of apoptosis after incubation with monocyte supernatant. After different incubation time periods, no change was found in bcl-2 protein expression in U937 and K562 cells, whereas in KG1, HL60, and especially in THP1 cells, a significant decrease could be noticed. On the other hand, there was an increase in bcl-2 expression in HL60-D3 cells. Bax protein expression, measured at the same time points, remained essentially unchanged in HL60-D3 cells, decreased significantly in U937, HL60, and THP1 cells and slightly in K562 cells, and increased significantly in KG1 cells. Also, the ratio bax/bcl-2 decreased in HL60D3, but especially in U937 and HL60 cells, increased slightly in THP1 and KG1 cells, and remained essentially unchanged in K562 cells. Rh-tumor necrosis factor-alpha (TNF-alpha), the main mediator of monocyte mediated cytotoxicity, induced apoptosis in U937, HL60, and THP1 cells, thereby showing changes in bcl-2 expression similar to those found for monocyte derived supernatants. We concluded that in human leukemic cell lines, there is no relation between either bcl-2 or bax protein expression or the ratio of both, and apoptosis mediated by monocyte derived supernatant or TNF-alpha.

Falls in the elderly: a prospective study of risk factors and risk profiles.

In this prospective study, the authors determined intrinsic risk factors for falls and recurrent falls and constructed a risk profile that indicated the relative contribution of each risk factor and also estimated the probabilities of falls and recurrent falls. In 1992, over a 28-week period, falls were recorded among 354 elderly subjects aged 70 years or over who were living in homes or apartments for the elderly in Amsterdam and the vicinity. During the study period, 251 falls were reported by 126 subjects (36%), and recurrent falls (> or =2 falls) were reported by 57 subjects (16%). Associations of falls and recurrent falls with potential risk factors were identified in logistic regression models. Mobility impairment regarding one or more of the tested items (i.e., impairment of balance, leg-extension strength, and gait) was associated with falls (adjusted odds ratio (OR) =2.6) and was strongly associated with recurrent falls (OR = 5.0). Dizziness upon standing was associated with falls (OR = 2.1) and recurrent falls (OR = 2.1). However, several risk factors were associated with recurrent falls only: history of stroke (OR = 3.4), poor mental state (OR = 2.4), and postural hypotension (OR = 2.0). The authors constructed a risk profile for recurrent falls that included the five risk factors mentioned above. Inclusion of all risk factors in the profile implied an 84% probability of recurrent falls over a period of 28 weeks, compared with 3% when no risk factor was present. The probability of recurrent falls ranged only from 11% to 29% when predicted by number of falls occurring in the previous year. Physical activity, use of high-risk medication, and the use of vitamin D3, which was randomly allocated to the participants, were not strongly related to either falls or recurrent falls. In conclusion, a large range of probabilities of falls, especially of recurrent falls, was estimated by the risk profiles, in which mobility impairment was the major risk factor. Recurrent fallers may therefore be especially amenable to prevention based on mobility improvement.