A simulated SNP experiment indicates a high risk of over-fitting and false positive results when a predictive multiple SNP model is established and tested within the same dataset

Association between single nucleotide polymorphism of DNA damage repair related genes and radiosensitivity in healthy individuals

Radiosensitivity in humans can influence radiation-induced normal tissue toxicity. As radiosensitivity has a genetic predisposition, we aimed to investigate the possible association between four single nucleotide polymorphism (SNP) sites and the radiosensitivity in healthy people. We genotyped four selected SNPs: TRIP12 (rs13018957), UIMC1 (rs1700490) and POLN (rs2022302), and analyzed the association between SNP and the radiosensitivity in healthy people. We distinguished radiosensitivity by chromosome aberration analysis in healthy individuals. Healthy donors were classified into three groups based on chromosomal aberrations: resistant, normal and sensitive. Using the normal group as a reference, the genotypes CT and CC of rs13018957 (CT: OR = 26.13; CC: OR = 15.97), AA of rs1700490 (OR = 32.22) and AG of rs2022302 (OR = 13.98) were risk factors for radiosensitivity. The outcomes of the present study suggest that four SNPs are associated with radiosensitivity. This study lends insights to the underlying mechanisms of radiosensitivity and improves our ability to identify radiosensitive individuals.

Predicting COVID-19 disease severity from SARS-CoV-2 spike protein sequence by mixed effects machine learning

Epidemiological studies show that COVID-19 variants-of-concern, like Delta and Omicron, pose different risks for severe disease, but they typically lack sequence-level information for the virus. Studies which do obtain viral genome sequences are generally limited in time, location, and population scope. Retrospective meta-analyses require time-consuming data extraction from heterogeneous formats and are limited to publicly available reports. Fortuitously, a subset of GISAID, the global SARS-CoV-2 sequence repository, includes "patient status" metadata that can indicate whether a sequence record is associated with mild or severe disease. While GISAID lacks data on comorbidities relevant to severity, such as obesity and chronic disease, it does include metadata for age and sex to use as additional attributes in modeling. With these caveats, previous efforts have demonstrated that genotype-patient status models can be fit to GISAID data, particularly when country-of-origin is used as an additional feature. But are these models robust and biologically meaningful? This paper shows that, in fact, temporal and geographic biases in sequences submitted to GISAID, as well as the evolving pandemic response, particularly reduction in severe disease due to vaccination, create complex issues for model development and interpretation. This paper poses a potential solution: efficient mixed effects machine learning using GPBoost, treating country as a random effect group. Training and validation using temporally split GISAID data and emerging Omicron variants demonstrates that GPBoost models are more predictive of the impact of spike protein mutations on patient outcomes than fixed effect XGBoost, LightGBM, random forests, and elastic net logistic regression models.

Single nucleotide polymorphisms might influence chemotherapy induced nausea in women with breast cancer

Background

Women receiving FEC (5 fluorouracil, epirubicin and cyclophosphamide) chemotherapy (CT) for breast cancer (BC) often experience side effects such as nausea and vomiting. Individual variations of side effects occur in patients despite similar cancer therapy. The purpose of this study was to investigate a possible genetic background as a predictor for individual variations in nausea induced by CT.

Methods

114 women were included in the study. All women received adjuvant CT for BC. Self-reported nausea and vomiting was recorded in a structured diary over ten days following treatment. Blood samples were collected before the treatment and used for the detection of 48 single nucleotide polymorphisms (SNPs) in 43 genes. SNPs from each individual woman were analyzed for their relation to the patient-reported frequency and intensity of nausea and vomiting.

Results

Eighty-four percent (n = 96) of the women reported acute or delayed nausea or combined nausea and vomiting during the ten days following CT. Three out of the forty-eight SNPs in the following genes: FAS/CD95, RB1/LPAR6 and CCL2 were found to be associated with a risk of nausea.

Conclusion

SNPs in the FAS/CD95, RB1/LPAR6 and CCL2 genes were found to be associated with nausea among women treated with adjuvant FEC for BC. SNPs analysis is fast and cost effective and can be done prior to any cancer therapy. The association between individual SNPs and severe side effects from FEC may contribute to a more personalized care of patients with BC.

DNA Methyltransferase Gene Polymorphisms for Prediction of Radiation-Induced Skin Fibrosis after Treatment of Breast Cancer: A Multifactorial Genetic Approach

Purpose

This study was conducted to investigate the role of four polymorphic variants of DNA methyltransferase genes as risk factors for radiation-induced fibrosis in breast cancer patients. We also assessed their ability to improve prediction accuracy when combined with mitochondrial haplogroup H, which we previously found to be independently associated with a lower hazard of radiation-induced fibrosis.

Materials and Methods

DNMT1 rs2228611,DNMT3A rs1550117,DNMT3A rs7581217, and DNMT3B rs2424908 were genotyped by real-time polymerase chain reaction in 286 Italian breast cancer patients who received radiotherapy after breast conserving surgery. Subcutaneous fibrosis was scored according to the Late Effects of Normal Tissue–Subjective Objective Management Analytical (LENT-SOMA) scale. The discriminative accuracy of genetic models was assessed by the area under the receiver operating characteristic curves (AUC).

Results

Kaplan-Meier curves showed significant differences among DNMT1 rs2228611 genotypes in the cumulative incidence of grade ≥ 2 subcutaneous fibrosis (log-rank test p-value= 0.018). Multivariate Cox regression analysis revealed DNMT1 rs2228611 as an independent protective factor for moderate to severe radiation-induced fibrosis (GG vs. AA; hazard ratio, 0.26; 95% confidence interval [CI], 0.10 to 0.71; p=0.009). Adding DNMT1 rs2228611 to haplogroup H increased the discrimination accuracy (AUC) of the model from 0.595 (95% CI, 0.536 to 0.653) to 0.655 (95% CI, 0.597 to 0.710).

Conclusion

DNMT1 rs2228611 may represent a determinant of radiation-induced fibrosis in breast cancer patients with promise for clinical usefulness in genetic-based predictive models.

Gender bias in individual radiosensitivity and the association with genetic polymorphic variations

PURPOSE

To assess the extent of variation in radiosensitivity between individuals, gender-related dissimilarity and impact on the association with single nucleotide polymorphisms (SNPs).

MATERIALS AND METHODS

Survival curves of 152 fibroblast cell strains derived from both gender were generated. Individual radiosensitivity was characterized by the surviving fraction at 2Gy (SF2). SNPs in 10 radiation responsive genes were genotyped by direct sequencing.

RESULTS

The wide variation in SF2 (0.12-0.50; mean=0.33) was significantly associated with 3 SNPs: TP53 G72C (P=0.007), XRCC1 G399A (P=0.002) and ATM G1853A (P=0.01). Females and males differed significantly in radiosensitivity (P=0.004) that impacted genetic association where only XRCC1 remained significant in both gender (P<0.05). Meanwhile, discordant association was observed for TP53 that was significant in females (P=0.012) and ATM that was significant in males (P=0.0006). When gender-specific SF2-mean (0.31 and 0.35 for females and males; respectively) was considered, further discordance was observed where XRCC1 turned out not to be associated with radiosensitivity in males (P>0.05).

CONCLUSIONS

Although the variation in individual radiosensitivity was associated with certain SNPs, gender bias for both endpoints was evident. Therefore, assessing the risk of radiation exposure in females and males should be considered separately in order to achieve the ultimate goal of personalized radiation medicine.