Broad Institute Genomic Services at AGBT 2017

Posters from this year's annual conference generated at Broad Genomics

 

Posters

Development of production scale cfDNA sequencing @ Broad Genomics

C. Cibulskis, V. Adalsteinsson, M. Fleharty, K. Larkin, J. Abreu, M. Cipicchio, J. Grimsby, B. Blumenstiel, M. DeFelice, T. Mullen, G. Ha, S. Freedman, K. Cibulskis, A. Choudhury, D. Stover, H. Parsons, N. Wagle, G. Gydush, S. Reed, D. Rotem, J. Rhoades, L. Lichtenstein, S. Gabriel, T. Golub, J. Boehm, G. Getz, J.C. Love, M. Meyerson, N. Lennon 

Genomic profiling of cell free tumor DNA, (Blood Biopsy), offers the potential to revolutionize cancer precision medicine.  As a proxy for tumor tissue profiling, successful blood biopsy analysis can help select appropriate patients for clinical trials, provide useful data for treatment monitoring, and discover genomic mechanisms of disease resistance.

A cloud-based genotyping analysis pipeline

G. Grant, J. Carey, K. Tibbetts

The use of genotyping arrays remains a cost-effective and valuable technique for validation of sequencing results and for running large-scale GWAS studies. At the Broad Institute we have developed a highly scalable, cloud-based genotyping array data analysis pipeline to facilitate our continued use of Illumina Sentrix genotyping arrays. This workflow follows on to and uses tools and strategies developed for our cloud-based Whole Genome Shotgun, (WGS), workflow.

Methods for detection of sample swaps, contaminants, and pedigrees

Y. Farjoun, M. Shand, J. Bloom, E. Banks

Sample identity in labs needs to be maintained from the input source to the data. A break in that chain of custody results in a sample-swap, corrupting the data and possibly creating false conclusions. We have developed lab procedures and fingerprinting software that can detect swaps and re-identify the sample. From the sequence data and independently measured genotypes we compute a score, (log odds), that informs whether the sample is correct or swapped.

Understanding low allele variant detection in heterogeneous samples, required read coverage and the utility of unique molecular indices (UMIs)

B. Blumenstiel, M. Fleharty, M. DeFelice, L. Green, J. Grimsby, Y. Farjoun, N. Lennon, S. Gabriel

As interest in tumor genomic profiling and circulating cell-free DNA sequencing continues to evolve, the ability to detect low allele frequency and rare genomic alterations in heterogeneous samples is becoming increasingly important. With clinically relevant variants often presenting with allele frequencies lower than 1%, deep sequencing coverage and highly sensitive and specific methods are needed for efficient screening and discovery of actionable alterations.  

Detection and depletion of bacteria contamination in saliva derived DNA samples to reduce costs and improve data utility for Human Whole Genome Sequencing

M. Costello, S. Freedman, Y. Farjoun, N. Lennon, S. Gabriel

Obtaining saliva for genomic analysis has become increasingly popular as collection methods are less invasive than blood draws or biopsies. However, a major drawback of saliva is the presence of non-human DNA contamination mainly due to oral bacteria. This bacterial contamination can require costly additional sequencing and can impact downstream data quality.   

Meeting the demand for long range sequencing: A scalable automated workflow for the 10X Chromium Whole Genome Platform

M. Cipicchio, C. Walsh, J. Walsh, J. Abreu, D. Perrin, T. DeSmet, S. Dodge, S. Gabriel

Long range sequencing is essential to understand the full variability of a whole genome, enabling phasing and revealing large structural variations. The Chromium Genome Solution from 10X Genomics enables users to recover long range sequence information from short read sequencing data.

Design and implementation of a high throughput fully automated sequencing based fingerprint assay

C. Walsh, J. Grimsby, Y. Farjoun, M. Shand

Currently, using a chip based genotyping assay, only 25% of the samples processed in the Genomics Platform at the Broad Institute are fingerprinted. This throughput is limited as a result of the associated workflow, sample input requirements, and elevated cost. To overcome these challenges, a sequencing-based fingerprinting assay was developed.