Novel computational analysis of large transcriptome datasets identifies sets of genes distinguishing chronic obstructive pulmonary disease from healthy lung samples

Abstract

Chronic obstructive pulmonary disease (COPD) kills over three million people worldwide every year. Despite its high global impact, the knowledge about the underlying molecular mechanisms is still limited. In this study, we aimed to extend the available knowledge by identifying a small set of COPD-associated genes. We analysed different publicly available gene expression datasets containing whole lung tissue (WLT) and airway epithelium (AE) samples from over 400 human subjects for differentially expressed genes (DEGs). We reduced the resulting sets of 436 and 663 DEGs using a novel computational approach that utilises a random depth-first search to identify genes which improve the distinction between COPD patients and controls along the first principle component of the data. Our method identified small sets of 10 and 15 genes in the WLT and AE, respectively. These sets of genes significantly (p < 10–20) distinguish COPD patients from controls with high fidelity. The final sets revealed novel genes like cysteine rich protein 1 (CRIP1) or secretoglobin family 3A member 2 (SCGB3A2) that may underlie fundamental molecular mechanisms of COPD in these tissues.

Development and reliability of the histological THROMBEX-classification rule for thrombotic emboli of acute ischemic stroke patients

“…Following the terms of cloud formations, the distribution of platelets is described as “stratus “, “cirrus “, or “cumulus “. Vivid pictures are published in [ 8 ]. The histological cross-sections were captured with 200-fold original magnification.…”

“…Thus, platelet aggregation was subdivided into “stratus”, “cirrus”, and “cumulus” [ 8 ]. Following the wording of the International Cloud Atlas, in a stratus distribution platelets occur as a merged layer with a fairly uniform appearance or in the form of ragged patches.…”

“…Histological features of emboli extracted from acute ischemic stroke patients were examined and correlated with the medical history and diagnostic workup to develop the first histological classification rule that can distinguish between cardiac and arteriosclerotic emboli in individual stroke patients. Its prediction accuracy is in the range of 70–78% [ 8 ].…”

“…The defined histological classification rule refers to these hypotheses and is suitable to differentiate between cardiac and arteriosclerotic emboli [ 8 ], which are the most frequent causes of proximal vessel occlusion of acute ischemic stroke patients [ 1 ].…”

“…We already published the validation of this classification rule [ 8 ]. However, sufficient reproducibility is also indispensable for its future application by different scientists or for repeated investigations performed by the same histopathologist.…”