Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2018 Jul 27;5:80-89.
doi: 10.1016/j.isci.2018.07.001. Epub 2018 Jul 6.

Histopathological Image QTL Discovery of Immune Infiltration Variants

Affiliations
Free PMC article

Histopathological Image QTL Discovery of Immune Infiltration Variants

Joseph D Barry et al. iScience. .
Free PMC article

Abstract

Genotype-to-phenotype association studies typically use macroscopic physiological measurements or molecular readouts as quantitative traits. There are comparatively few suitable quantitative traits available between cell and tissue length scales, a limitation that hinders our ability to identify variants affecting phenotype at many clinically informative levels. Here we show that quantitative image features, automatically extracted from histopathological imaging data, can be used for image quantitative trait loci (iQTLs) mapping and variant discovery. Using thyroid pathology images, clinical metadata, and genomics data from the Genotype-Tissue Expression (GTEx) project, we establish and validate a quantitative imaging biomarker for immune cell infiltration. A total of 100,215 variants were selected for iQTL profiling and tested for genotype-phenotype associations with our quantitative imaging biomarker. Significant associations were found in HDAC9 and TXNDC5. We validated the TXNDC5 association using GTEx cis-expression QTL data and an independent hypothyroidism dataset from the Electronic Medical Records and Genomics network.

Keywords: Association Analysis; Bioinformatics; Computational Bioinformatics; Pathology.

Figures

None
Figure 1
Figure 1
Thyroid Image Processing and Establishment of a Quantitative Imaging Biomarker for Immune Cell Infiltration (A) Digital pathology slide for thyroid sample GTEX-11NV4. Raw: before image processing. Blurred: post Gaussian convolution. Segmented: tissue piece masks after adaptive thresholding. (B) Heatmap of 117 log2 transformed and standardized Haralick image features against 341 thyroid samples. (C) PC1 versus PC2 from a PCA of the image feature matrix. Blue points indicate patients with Hashimoto thyroiditis, as identified from pathology notes. Circles indicate females and triangles males. (D) ROC showing biomarker performance of PC2 for predicting HT. (E) Top row, five thyroid samples with highest values of image PC2. Bottom row, local image PC2 signal. Bright regions correspond to high local image PC2. Image brightness has been rescaled to aid visualization.
Figure 2
Figure 2
Integration of Imaging Data with Gene Expression Analyses Verifies that Image PC2 Is Highly Associated with Immune Cell Infiltration (A) A QQ plot of p values (PG) from the regression analysis of image PC1 and PC2 against thyroid gene expression for 23,993 genes. (B) Correlation of image PC2 with −log10(PC) from the CIBERSORT analysis for samples with PC<0.5. Blue points are samples for which Hashimoto thyroiditis was indicated in GTEx pathology notes. (C) Frequencies of immune cell types reported from CIBERSORT for samples with PC<0.1. Cell types with an average frequency of 5% or more are shown.
Figure 3
Figure 3
An Image QTL Analysis Finds Associations between Genomic Variants and Our Image PC2 Biomarker for Thyroid Autoimmune Disease (A) Selection of 1,380 candidate genes (blue points) based on their positive fold-change (log2(FC) > 0.5) and significant differential expression (−log10(Padj) > 7) in GTEx samples with Hashimoto thyroiditis (HT) phenotypes. (B) A QQ plot showing expected vs observed p values from image QTL fits of 100,215 candidate SNPs residing in the selected genes highlighted blue in panel A. (C and D) −log10(P) vs genomic coordinates for GTEx iQTLs (top panels) and eMERGE variant association with HT (bottom panels) for all tested SNPs in HDAC9 (C) and TXNDC5 (D). Blue vertical lines indicate the locations of the most significant SNP for each gene after multiple testing correction using the IHW method described in the main text. Plot ranges are mapped to the start and end positions of the genes, as defined by GTEx-consortium transcript data. Horizontal green bars are of length 50 kb.

Similar articles

See all similar articles

Cited by 5 articles

References

    1. Aguet F., Brown A.A., Castel S.E., Davis J.R., He Y., Jo B., Mohammadi P., Park Y., Parsana P., Segrè A.V. Genetic effects on gene expression across human tissues. Nature. 2017;550:204–213. - PubMed
    1. Anderson C.A., Boucher G., Lees C.W., Franke A., D'Amato M., Taylor K.D., Lee J.C., Goyette P., Imielinski M., Latiano A. Meta-analysis identifies 29 additional ulcerative colitis risk loci, increasing the number of confirmed associations to 47. Nat. Genet. 2011;43:246–252. - PMC - PubMed
    1. Anila K., Nayak N., Jayasree K. Cytomorphologic spectrum of lymphocytic thyroiditis and correlation between cytological grading and biochemical parameters. J. Cytol. 2016;33:145–149. - PMC - PubMed
    1. Beck A.H., Sangoi A.R., Leung S., Marinelli R.J., Nielsen T.O., van de Vijver M.J., West R.B., van de Rijn M., Koller D. Systematic analysis of breast cancer morphology uncovers stromal features associated with survival. Sci. Transl. Med. 2011;3:108ra113. - PubMed
    1. Beck T., Hastings R.K., Gollapudi S., Free R.C., Brookes A.J. GWAS Central: a comprehensive resource for the comparison and interrogation of genome-wide association studies. Eur. J. Hum. Genet. 2014;22:949–952. - PMC - PubMed

LinkOut - more resources

Feedback