Please read the manuscript and the userguide before using.
Github: mskcc/facets
DOI: https://doi.org/10.1093/nar/gkw520
Citation: Ronglai Shen, Venkatraman E. Seshan; FACETS: allele-specific copy number and clonal heterogeneity analysis tool for high-throughput DNA sequencing, Nucleic Acids Research, Volume 44, Issue 16, 19 September 2016, Pages e131, https://doi.org/10.1093/nar/gkw520
This repository contains an implementation of FACETS, an algorithm to estimate fraction and allele specific copy number from tumor/normal sequencing by Ronglai Shen and Venkatraman E. Seshan from Memorial Sloan-Kettering Cancer Center's Department of Epidemiology and Biostatistics, for FireCloud. FACETS is run on tumor-normal pairs to determine the purity and ploidy of a tumor sample, as well as infer allele-specific copy number.
To best use this repository, please:
- Read the manuscript and the userguide
- Consider the coverage of your samples, especially when thinking of the ndepth (minimum read depth in normal) parameter.
- Check the emflags to see if your sample is noisy
- See how stable your called solution(s) are with the .facets_iterations.txt and pdf outputs
- Observing NA purity values? See this Github issue and also this one
- Read and search the issues, both open and closed, on Github to learn more about the method
- Read the help text on each function called in facets.R
- Inspect each sample closely
Docker image: vanallenlab/facets
FireCloud method: vanallenlab/facets
Read depth and allele counts at sites of common single nucleotide variant are observed in both the tumor and normal bams based on a provided VCF.
Inputs:
- Pair name
- Normal bam and corresponding index
- Tumor bam and corresponding index
- VCF and corresponding index for common variants
- Minimum mapping quality
- Minimum based quality
- Minimum read depth in normal and tumor
- Number of pseudo SNPs
Read the pileup documentation for more details of each input. The default VCF of common variants in the FireCloud method configuration is human 9606 b151 GRCh37, if you do not have access please download the appropriate VCF and host in a bucket that you have access to.
Outputs:
- Pileup in gunzip format. Read more in the userguide.
Estimates fraction and allele specific copy number from paired tumor/normal sequencing. As a result, also estimates purity and ploidy of a given tumor sample. This implementation will run FACETS 10 times across different seeds to observe how stable the inferred purity and ploidy values are for a given pair, the seed with a purity closest to the median purity value is selected. If you would prefer to not iterate and use a specific seed, set seed_iterations to 1 and specify your preferred seed with seed_initial.
Inputs:
- Pair name
- Pileup
- Minimum read depth in normal
- Critical value (cval), the threshold for determining if a change exists
- Maximum number of expectation maximization iterations (maxiter)
- Initial seed to set R to
- Number of seeds to test
Outputs:
- Genome segments plot, displaying copy and log ratios across chromosomes
- Diagnostics plot, displaying the segment summaries
- Iterations plot, displys ploidy and purity for all iterations performed across seeds
- Copy number cellular fraction, total and minor allele counts for all segments
- Estimated purity and ploidy
- Flags and warnings generated by FACETS
- Seed value closest to median purity
- Number of seeds that resulted in purity values equaling NA
Infers whole genome doubling based on Bielski CM, Zehir A, Penson AV, et al. Genome doubling shapes the evolution and prognosis of advanced cancers. Calculates major copy number (MCN) estimate based on total copy number (TCN) estimate and minor copy number (LCN) estimate from FACETS and calls whole-genome doubling if the average MCN across the autosomal genome is greater than 2. In cases that LCN is equal to NA, a value of 0 is used.
Inputs:
- Copy number cellular fractions
Outputs:
- Fraction major copy number greater than 2
- Putative whole genome doubling, if the fraction of MCN across the autosomal genome is greater than 0.5
Infers the percentage of the genome altered by copy number alterations (differences from normal ploidy). Commonly used in a variety of papers, (example here). This is done by taking all the segments that do not match diploid copy number (2 for autosomes and 1 for sex chromosomes) and computing their sizes, then dividing by the total size of all the segments. In cases that total copy number is equal to NA, a value of 0 is used, although this should be impossible.
Inputs:
- Copy number cellular fractions
Outputs:
- Fraction of genome altered
- Ronglai Shen, Venkatraman E. Seshan; FACETS: allele-specific copy number and clonal heterogeneity analysis tool for high-throughput DNA sequencing, Nucleic Acids Research, Volume 44, Issue 16, 19 September 2016, Pages e131, https://doi.org/10.1093/nar/gkw520
- Bielski CM, Zehir A, Penson AV, et al. Genome doubling shapes the evolution and prognosis of advanced cancers
- FACETS Github issue, "filtering and interpreting results"
- FACETS Github issue, "About parameters setting for targeted panel
- FACETS Github issue, "About snp pileup, psuedo snps