20190725_-_2019-36253_-_Statement_of_Work_(SOW).pdf

STATEMENT OF WORK (SOW)

Title: Broad Spectrum Human Papillomavirus Genotyping using Single Molecule Whole Genome Sequencing- Extended development of the bioinformatics pipeline for genotyping of human papillomavirus including calling new and known HPV variants and identification of HPV integration status

SECTION 1 – BACKGROUND

Human Papillomavirus (HPV) testing is essential for HPV surveillance studies of public health importance and to evaluate the burden of HPV associated cancers like cervical cancer and the impact of HPV vaccination. There are over 200 HPV genotypes but current methods can detect only approximately 37 genotypes. There is no single laboratory test that can detect all HPV genotypes in a given sample, leaving an incomplete picture in the population of the genotypes, variants, and physical status of HPV, the most sexually transmitted infectious agent. Since HPV cannot be cultured, its detection is totally dependent on molecular tools. The unmet need of developing a broad spectrum HPV genotyping test depends essentially on advanced molecular detection (AMD) technologies and bioinformatics. On this rationale, CDC HPV laboratory at DHCPP-CVDB began developing a validated baited Single Molecule Sequencing HPV genotyping test (baited HPV-SMS genotyping test) that would allow for the detection of all known HPV genotypes, and also new HPV types which is not possible with the current testing. The key steps in the development of this test will essentially involve those outlined for the human exome sequencing using Illumina HiSeq and Agilent SureSelectXT2Target Enrichment System, but the target enrichment system consisting of biotinylated RNA library “baits” will be customized for the full length genome (approximately 8 kb) of all known HPV types. Representing long or nearly full-length sequences of all possible HPV genomes in a sample is critical to the specificity of type-assignment, and potentially enabling the determination of variants within each genotype, critical information that is not obtained by the current genotyping methods. This new baited HPV-SMS genotyping test would subsequently be evaluated for its performance in comparison to current method(s), and compatibility with sample collection methods including formalin-fixed paraffin embedded (FFPE) samples. The HPV-SMS genotyping test will also be evaluated for performance metrics in terms of accuracy, precision, sensitivity, specificity, reproducibility and reportable range through series of studies utilizing WHO reference materials and clinical specimens. A validated HPV-SMS genotyping test having quality performance metrics will then be piloted to HPV surveillance studies in evaluating the impact of HPV vaccination and potential for type-replacement, re-testing of HPV-negative cancers of different anatomical sites, and creating a large-scale HPV whole genome sequence database using total nucleic acid (TNA) extracts from cervical specimens linked to NIH Early Detection Research Network (EDRN). The CDC-HPV laboratory has specimens from a number of HPV vaccination studies and cancer registries to conduct these studies using the HPV- SMS genotyping test. The CDC HPV laboratory is currently one of two global reference laboratories, and these proposed studies using AMD technologies and bioinformatics are essential to maintaining the global reference laboratory status of the CDC-HPV lab.

CDC HPV lab, in collaboration with Agilent Technologies, completed design and production of RNA library baits (overlapping 120-mer) representing all known HPV genomes. These HPV specific RNA baits were used to capture HPV specific sequences in a pilot study using 32 samples consisting of clinical, cell lines and WHO reference materials. To capture HPV specific sequences, sample

DNAs were sheared, adaptor modified, barcoded, and pooled (up to 16 samples) prior to hybridization to HPV specific RNA library baits in excess. After hybridization, background DNA was washed away, and the target HPV DNA was eluted, amplified (under reduced cycles) and pair-end sequenced by Illumina platform. In addition to this initial Illumina sequence data from 32 piloted samples, CVDB also generated HPV whole genome sequence data using (1) pooled HPV plasmids (9 HPV vaccine types) and cell lines to determine the reproducibility and limit of detection of the whole genome sequencing assay, (2) DNA extracted from external genital swabs (50 samples), and (3) using hundreds of total nucleic acid extracts from EDRN biorepository completed building a largescale HPV whole genome sequence database for the development of bioinformatics tools for automatic determination of HPV genotype, variant and integration status in clinical specimens from epidemiologic studies. Current laboratory work is focused on optimizing the DNA library preparation and bioinformatics tools for whole genome sequencing of HPV in FFPE samples following target enrichment, and incorporation of molecular barcodes/dual indexing to address the issues of duplicates and index hoping.

CVDB has no personnel with bioinformatics expertise in the analysis of next generation sequence (NGS) data for processing of the sequence data sets from a variety of sample sources to enable HPV type and variant determination and assessment of physical status of the dominant types in a given samples. The Government issued a contract for this work starting in September 2016. Through this contract, the following has been accomplished.

1. Developed novel algorithms and software tools that expanded the earlier developed broad-spectrum genotyping pipeline. The algorithm allows for automatic identification of known and new HPV types present as dominant or non-dominant HPV agents in epidemiological samples. Parameters and thresholds of the pipeline for automated HPV genotyping were optimized with the application of support vector machine (SVM) learning algorithm.

2. Validated automated HPV genotyping results obtained with the newly developed pipeline.

The results were compared with the non-automated HPV genotyping results obtained by the commercially available software for read mapping.

3. Determined the sensitivity and specificity of automated bioinformatics pipeline based on SVM for HPV genotyping.

4. Completed the analysis of largescale HPV whole genome sequence data from EDRN, and results are being delivered to CDC.

5. Undertook revision of the manuscript titled “Human Papillomavirus Genotyping by SVM Classification of NGS Data” for publication in a scientific journal.

6. Tested performance of the bioinformatics pipeline for calling HPV types using whole genome sequence data generated using Agilent XT baits and MiSeq platform.

7. Contributed to the conceptual framework to expand the genotyping pipeline for incorporating algorithms and software tools for automated calling of known and novel variants, and detection of HPV integration into the human genome

SECTION 2 – SCOPE OF WORK

The Government continues to require bioinformatics services for another year to expand the application of pipeline for determination of variant and integration status, in addition to genotyping HPV. The already conducted research opened new horizons in terms of additional avenues of research that would make the outcomes more precise with respect to changing technology and new volumes of accumulated data on HPV whole genome sequencing at CDC.

SECTION 3 – SCOPE OF WORK

As part of this SOW, the Contractor shall provide expertise and direction to develop work-flows, and novel algorithms to automate determination of HPV types, variants and its physical status (integration of HPV genome into human genome) using whole genome sequence data from multiple sources including challenging materials like FFPE; generating multiple levels of information from the sequence data which is not dependent on the library preparation or sequencing platforms (HiSeq vs MiSeq). Work flows and algorithm development shall take into consideration of evaluating multiple parts of the whole genome and single and multiple infections of HPV types in a clinical sample. A main thrust and outcome of this contract shall be to continue development of the customized bioinformatics pipeline that automates determination of genotype, variant and integration status of HPV in a given sample. The Government expects several bioinformatics challenges in developing reliable algorithms for HPV genotyping, variant assignment and determination of integration status in a single pipeline.

Important part of the work associated with this requirement is expanding the application of bioinformatics pipeline for determination of HPV variant and integration status, and making the pipeline application independent of sequencing platforms.

SECTION 4 – TECHNICAL REQUIREMENTS

The Contractor shall perform the following tasks.

1. Develop novel algorithms and software tools that shall increase the precision of the current broad-spectrum genotyping pipeline using information on NGS short reads incorporating molecular barcodes and dual index. The new algorithm extensions shall focus on automate identification of the known and new HPV types and variants present as dominant or non-dominant HPV agents in epidemiological samples. Parameters and thresholds of the pipeline for automated HPV variant determination shall be further optimized with respect to possible incomplete coverage of HPV genome by short reads. Particular attention shall be given to detecting SNPs in E7 gene involved in carcinogenesis.

2. Validate the automated HPV genotyping and variant determination by the new version of the pipeline. Automated pipeline results shall be compared with the results from non-automated commercially available software for read mapping.

3. Improve the sensitivity and specificity of automated bioinformatics pipeline results based on machine learning (SVM) for HPV genotyping and variant determination.

4. Develop the HPV variant database with entries added with help of new algorithms for automated calling of known and novel HPV variants, lineages and sub-lineages.

5. Develop new algorithms and software tools for expanding the broad-spectrum genotyping pipeline for detecting HPV integration into human genome. The focus shall be on characterization of the HPV variants as well as identification of HPV integration status.

SECTION 5 – DELIVERABLES

The Contractor shall provide the following deliverables according to the schedule below.

Deliverable Delivery Date Delivery Method Recipient

Quarterly Report quarterly from the date of award (e.g.

mid-December 2019, mid-March 2020, mid- June 2020, and mid- September 2020) electronically via email or in-person presentation

Dr. Mangalathu Rajeevan (mor4@cdc.gov)

SECTION 6 – GOVERNMENT-FURNISHED MATERIALS (GFM)

No Government-furnished Materials (GFM) will be provided under the anticipated contract.

SECTION 7 – PERIOD OF PERFORMANCE

The anticipated Period of Performance is twelve months from date of award. Specific Period of Performance dates are anticipated to be 1 September 2019 through 31 August 2020, although the contract may be awarded before these dates.

SECTION 8 – PLACE OF PERFORMANCE

Work performed under the anticipated contract will be performed at Contractor’s facility.

SECTION 9 – POINTS OF CONTACT

Government and Contractor Points of Contact will be provided at the time of contract award. All inquiries about this SOW shall be directed to the following Contract Specialist.

Lee Catherine Kennedy vhg0@cdc.gov mailto:mor4@cdc.gov mailto:vhg0@cdc.gov