Introduction: Retinoblastoma (Rb) is an intraocular malignant tumour of early childhood, considered the most robust clinical model of genetic predisposition to develop cancer. Previous transcriptomic studies in Rb are based on differential analysis of expression data, ignoring the detection scores that can be obtained from high throughput technologies. The detection score allows to obtain profiles of present/absent mRNAs across samples without the need of a control sample. In this study, 8 primary cultured tumors and Rb cell line Y79 were analyzed using the detection scores, with the aim to discover what is similar and what is different among them in terms of mRNA detection.

Methodology: Tumors were obtained from enucleated eyes of Rb patients treated at Hospital de Pediatría, CMN SXXI, Instituto Mexicano Del Seguro Social (IMSS) and Hospital Infantil de México Federico Gómez, Secretaría de Salud in Mexico City. Tumor tissues were cultured for a week and total RNA extracted. We used Affymetrix Human Transcriptome Array 2.0 able to detect many isoforms. Pre-processing included normalization with RMA (Average Robust Multiarray) and DABG (Detection Above Background) to calculate the detection score for each exon as present or absent. We carried out subsequent analysis to determine the detection score of each transcript and each gene in two corresponding datasets. For each of these detection datasets, hierarchical clustering (HC) was performed producing two color detection maps for present and absent elements.

Results: We obtained detection scores for 21,986 genes and 59,629 annotated transcripts (isoforms). In the detection map of genes obtained with HC, we discovered a central core of 3,391 mRNAs cluster shared by all the samples, while in the detection map of transcripts we discovered a 12,799 mRNAs cluster shared by all the samples. Absent mRNAs cluster were also shared across the samples with 9,005 in the genes and 18,651 in the transcripts detection maps. Notably, variability accounted for a cluster of 9,590 and 28,179 across the genes and transcripts detection maps, respectively. Furthermore, we determined the Rb transcriptomic size by counting the mRNAs in each sample and calculating the average across the datasets with 8,238 and 28,396 mRNA detected, 13,747 and 31,232 not detected corresponding to gene or transcript detection maps. When Y79 was introduced in the datasets, these numbers change considerably, and the corresponding detection maps show mRNA clusters specific for the cell line, but not present in the primary tissues.

Conclusion: Rb samples shared a central mRNA cluster of detected elements and a cluster of not detected elements, despite tumor heterogeneity. This systems level analysis brings a coherent and easy way for biological interpretation of high throughput data.

Note: This abstract was not presented at the meeting.

Citation Format: Diana E. Alvarez Suarez, Hugo Tovar, Enrique Hernández, Manuela Orjuela, Lourdes Cabrera, Claudia Hernández, Daphne García, Adriana Hernández, Verónica Ponce. Transcriptomic analysis of primary retinoblastoma with affymetrix HTA2 chip using a detection score reveals a central core of mRNA shared by all the samples [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 5143.