1. |
|
|
2. |
|
|
3. |
- Razzaq, M., et al.
(author)
-
Explainable Artificial Neural Network for Recurrent Venous Thromboembolism Based on Plasma Proteomics
- 2021
-
In: Computational Methods in Systems Biology19th International Conference, CMSB 2021, Bordeaux, France, September 22–24, 2021, Proceedings. - Cham : Springer Science and Business Media Deutschland GmbH. ; , s. 108-121
-
Conference paper (peer-reviewed)abstract
- Venous thromboembolism (VTE) is the third most common cardiovascular disease, affecting ∼ 1,000,000 individuals each year in Europe. VTE is characterized by an annual recurrent rate of ∼ 6%, and ∼ 30% of patients with unprovoked VTE will face a recurrent event after a six-month course of anticoagulant treatment. Even if guidelines recommend life-long treatment for these patients, about ∼ 70% of them will never experience a recurrence and will receive unnecessary lifelong anti-coagulation that is associated with increased risk of bleeding and is highly costly for the society. There is then urgent need to identify biomarkers that could distinguish VTE patients with high risk of recurrence from low-risk patients. Capitalizing on a sample of 913 patients followed up for the risk of VTE recurrence during a median of ∼ 10 years and profiled for 376 plasma proteomic antibodies, we here develop an artificial neural network (ANN) based strategy to identify a proteomic signature that helps discriminating patients at low and high risk of recurrence. In a first stage, we implemented a Repeated Editing Nearest Neighbors algorithm to select a homogeneous sub-sample of VTE patients. This sub-sample was then split in a training and a testing sets. The former was used for training our ANN, the latter for testing its discriminatory properties. In the testing dataset, our ANN led to an accuracy of 0.86 that compared to an accuracy of 0.79 as provided by a random forest classifier. We then applied a Deep Learning Important FeaTures (DeepLIFT) – based approach to identify the variables that contribute the most to the ANN predictions. In addition to sex, the proposed DeepLIFT strategy identified 6 important proteins (DDX1, HTRA3, LRG1, MAST2, NFATC4 and STXBP5) whose exact roles in the etiology of VTE recurrence now deserve further experimental validations.
|
|
4. |
|
|
5. |
- Bruzelius, Maria, et al.
(author)
-
PDGFB, a new candidate plasma biomarker for venous thromboembolism : results from the VEREMA affinity proteomics study
- 2016
-
In: Blood. - : American Society of Hematology. - 0006-4971 .- 1528-0020. ; 128:23, s. E59-E66
-
Journal article (peer-reviewed)abstract
- There is a clear clinical need for high-specificity plasma biomarkers for predicting risk of venous thromboembolism (VTE), but thus far, such markers have remained elusive. Utilizing affinity reagents from the Human Protein Atlas project and multiplexed immuoassays, we extensively analyzed plasma samples from 2 individual studies to identify candidate protein markers associated with VTE risk. We screened plasma samples from 88 VTE cases and 85 matched controls, collected as part of the Swedish Venous Thromboembolism Biomarker Study, using suspension bead arrays composed of 755 antibodies targeting 408 candidate proteins. We identified significant associations between VTE occurrence and plasma levels of human immunodeficiency virus type I enhancer binding protein 1 (HIVEP1), von Willebrand factor (VWF), glutathione peroxidase 3 (GPX3), and platelet-derived growth factor beta (PDGFB). For replication, we profiled plasma samples of 580 cases and 589 controls from the French FARIVE study. These results confirmed the association of VWF and PDGFB with VTE after correction for multiple testing, whereas only weak trends were observed for HIVEP1 and GPX3. Although plasma levels of VWF and PDGFB correlated modestly (rho similar to 0.30) with each other, they were independently associated with VTE risk in a joint model in FARIVE (VWF P < .001; PDGFB P = .002). PDGF. was verified as the target of the capture antibody by immunocapture mass spectrometry and sandwich enzyme-linked immunosorbent assay. In conclusion, we demonstrate that high-throughput affinity plasma proteomic profiling is a valuable research strategy to identify potential candidate biomarkers for thrombosis-related disorders, and our study suggests a novel association of PDGFB plasma levels with VTE.
|
|
6. |
- Bruzelius, M., et al.
(author)
-
Predicting venous thrombosis in women using a combination of genetic markers and clinical risk factors
- 2015
-
In: Journal of Thrombosis and Haemostasis. - : Elsevier BV. - 1538-7933 .- 1538-7836. ; 13:2, s. 219-227
-
Journal article (peer-reviewed)abstract
- BackgroundFamily history of venous thromboembolism (VTE) has been suggested to be more useful in risk assessment than thrombophilia testing. ObjectivesWe investigated established genetic susceptibility variants for association with VTE and evaluated a genetic risk score in isolation and combined with known trigger factors, including family history of VTE. Patients/MethodA total of 18 single nucleotide polymorphisms (SNPs) selected from the literature were genotyped in 2835 women participating in a Swedish nationwide case-control study (the ThromboEmbolism Hormone Study [TEHS]). Association with VTE was assessed by odds ratios (ORs) with 95% confidence interval (CI) using logistic regression. Clinical and genetic predictors that contributed significantly to the fit of the logistic regression model were included in the prediction models. SNP-SNP interactions were investigated and incorporated into the models if found significant. Risk scores were evaluated by calculating the area under the receiver-operating characteristics curve (AUC). ResultsSeven SNPs (F5 rs6025, F2 rs1799963, ABO rs514659, FGG rs2066865, F11 rs2289252, PROC rs1799810 and KNG1 rs710446) with four SNP-SNP interactions contributed to the genetic risk score for VTE, with an AUC of 0.66 (95% CI, 0.64-0.68). After adding clinical risk factors, which included family history of VTE, the AUC reached 0.84 (95% CI, 0.82-0.85). The goodness of fit of the genetic and combined scores improved when significant SNP-SNP interaction terms were included. ConclusionPrediction of VTE in high-risk individuals was more accurate when a combination of clinical and genetic predictors with SNP-SNP interactions was included in a risk score.
|
|
7. |
|
|
8. |
|
|
9. |
|
|
10. |
|
|