mskcc-omics-workflows · price0416 · Aug 6, 2026
diff --git a/containers/facets_tools_2n/3.0.0/Dockerfile b/containers/facets_tools_2n/3.0.0/Dockerfile
@@ -0,0 +1,199 @@
+FROM ghcr.io/mskcc-omics-workflows/r_tidyverse:3.6.3 AS runtime_base
+
+LABEL org.opencontainers.image.vendor="MSKCC-OMICS-WORKFLOWS" \
+      org.opencontainers.image.authors="Adam Price (price2@mskcc.org), Anne Marie Noronha (noronhaa@mskcc.org), Yixiao Gong (gongy@mskcc.org), Nikhil Kumar (kumarn1@mskcc.org), Philip Jonsson (jonssonp@mskcc.org), Ryan Ptashkin (ptashkir@mskcc.org)" \
+      org.opencontainers.image.created="2026-08-06T00:00:00Z" \
+      org.opencontainers.image.licenses="ACADEMIC SOFTWARE LICENSE AGREEMENT" \
+      org.opencontainers.image.version="3.0.0" \
+      imageprivacy="False" \
+      org.opencontainers.image.source="https://github.com/mskcc-omics-workflows/containers/containers/facets_tools_2n/" \
+      org.opencontainers.image.url="https://github.com/mskcc/facets-suite-2n" \
+      org.opencontainers.image.title="Facets Tools 2N" \
+      org.opencontainers.image.description="FACETS copy number suite for the unmatched-normal (2N) workflow: facets-suite-2n 3.0.0, facets2n, facets, snp-pileup and facets-preview."
+
+ENV DEBIAN_FRONTEND=noninteractive \
+    HTSLIB_VERSION=1.5 \
+    HTSTOOLS_VERSION=0.1.1 \
+    FACETS_SUITE_VERSION=3.0.0 \
+    FACETS_VERSION=0.5.14 \
+    PCTGCDATA=0.3.0 \
+    DATA_TABLE_VERSION=1.17.4 \
+    SNP_PILEUP=/usr/bin/snp-pileup \
+    PYTHONNOUSERSITE=set \
+    FACETS_OVERRIDE_EXITCODE=set \
+    LANG=C.UTF-8
+
+# facets2n, facets-suite-2n and facets-preview have no release tags at these
+# versions, so each is pinned to an explicit commit rather than a branch tip to
+# keep rebuilds reproducible. FACETS2N_SHA carries the flat-genome dipLogR guard
+# in findDiploidLogR.
+ENV FACETS2N_SHA=34016dcbbfdc8dc5583388a782bd442a545a683c \
+    FACETS_SUITE_2N_SHA=d66ad19499181b1f8c557619e5b7e0652c94ea5c \
+    FACETS_PREVIEW_SHA=a78fe1708665c17223893c63084805bb2639217d
+
+SHELL ["/bin/bash", "-o", "pipefail", "-c"]
+
+RUN apt-get update -qq \
+    && apt-get install -y --no-install-recommends \
+        ca-certificates \
+        git \
+        python3 \
+        python3-pip \
+        wget \
+        xvfb \
+    && apt-get clean \
+    && rm -rf /var/lib/apt/lists/*
+
+FROM runtime_base AS build_base
+
+# The r_tidyverse base already carries Cairo, argparse, gridExtra, binom, egg,
+# shiny, shinyWidgets, shinyjs, rhandsontable, doParallel, configr, R.utils,
+# shinyFiles, here, dplyr, httr, jsonlite, data.table, lifecycle, bit64,
+# lubridate, digest, glue, stringr, bslib and rtracklayer, so only these four
+# need installing here.
+#
+# bslib is not pinned. It is only ever a transitive dependency of shiny and
+# shinyWidgets -- no facets code references it. The base's shiny 1.7.3 needs
+# bslib (>= 0.3.0), which the base's 0.4.1 satisfies.
+#
+# reshape2 is not required by any of the three packages' DESCRIPTIONs, but
+# facets-preview ships a top-level arranging_facets.R that calls
+# library(reshape2), and that script is copied into the image below. It is kept
+# so that script stays runnable.
+#
+# data.table is upgraded past the base's snapshot version. facets-suite-2n only
+# declares data.table (>= 1.11.8), but it leans on it heavily, so this pins the
+# version the 2N pipeline was actually validated against rather than inheriting
+# whatever the base's CRAN snapshot happens to carry. The final stage asserts
+# the exact version, so a base bump cannot silently move it.
+RUN apt-get update -qq \
+    && apt-get install -y --no-install-recommends \
+        build-essential \
+        bzip2 \
+        gfortran \
+        libbz2-dev \
+        libc6-dev \
+        libcurl4-openssl-dev \
+        liblzma-dev \
+        libxt-dev \
+        python3-dev \
+        tar \
+        zlib1g-dev \
+    && R -e "options(repos = Sys.getenv('CRAN_MIRROR')); \
+    install.packages(c('plyr','diptest','DT','reshape2'), type = 'source')" \
+    && R -e "remotes::install_version('data.table', version = Sys.getenv('DATA_TABLE_VERSION'), \
+    repos = 'https://cloud.r-project.org', upgrade = 'never')" \
+    && apt-get clean \
+    && rm -rf /var/lib/apt/lists/*
+
+FROM build_base AS download_and_unpack
+
+WORKDIR /tmp
+
+RUN wget -q "https://github.com/samtools/htslib/releases/download/${HTSLIB_VERSION}/htslib-${HTSLIB_VERSION}.tar.bz2" \
+    && wget -q "https://github.com/mskcc/htstools/archive/snp_pileup_${HTSTOOLS_VERSION}.tar.gz" \
+    && wget -q "https://github.com/mskcc/facets/archive/v${FACETS_VERSION}.tar.gz" -O "facets-v${FACETS_VERSION}.tar.gz" \
+    && wget -q "https://github.com/mskcc/pctGCdata/archive/v${PCTGCDATA}.tar.gz" -O "pctGCdata-v${PCTGCDATA}.tar.gz" \
+    && tar xjf "htslib-${HTSLIB_VERSION}.tar.bz2" \
+    && tar xzf "snp_pileup_${HTSTOOLS_VERSION}.tar.gz" \
+    && tar xzf "facets-v${FACETS_VERSION}.tar.gz" \
+    && tar xzf "pctGCdata-v${PCTGCDATA}.tar.gz" \
+    && git clone --quiet https://github.com/mskcc/facets2n.git facets2n \
+    && git -C facets2n checkout --quiet "${FACETS2N_SHA}" \
+    && git clone --quiet https://github.com/mskcc/facets-suite-2n.git "facets-suite-${FACETS_SUITE_VERSION}" \
+    && git -C "facets-suite-${FACETS_SUITE_VERSION}" checkout --quiet "${FACETS_SUITE_2N_SHA}" \
+    && git clone --quiet https://github.com/mskcc/facets-preview.git facets-preview \
+    && git -C facets-preview checkout --quiet "${FACETS_PREVIEW_SHA}"
+
+FROM download_and_unpack AS htslib_builder
+WORKDIR "/tmp/htslib-${HTSLIB_VERSION}"
+RUN ./configure && make && make install && cp libhts.so* /usr/lib
+
+FROM htslib_builder AS htstools_builder
+WORKDIR "/tmp/htstools-snp_pileup_${HTSTOOLS_VERSION}"
+RUN g++ -std=c++11 snp-pileup.cpp -lhts -o snp-pileup \
+    && g++ -std=c++11 ppflag-fixer.cpp -lhts -o ppflag-fixer
+
+FROM htstools_builder AS pctgcdata_builder
+WORKDIR "/tmp/pctGCdata-${PCTGCDATA}"
+RUN R CMD INSTALL .
+
+FROM pctgcdata_builder AS facets_builder
+WORKDIR "/tmp/facets-${FACETS_VERSION}"
+RUN R CMD INSTALL .
+
+FROM facets_builder AS facets2n_builder
+WORKDIR /tmp/facets2n
+RUN R CMD INSTALL .
+
+FROM facets2n_builder AS facets_suite_builder
+WORKDIR "/tmp/facets-suite-${FACETS_SUITE_VERSION}"
+RUN R CMD INSTALL .
+
+FROM facets_suite_builder AS facets_preview_builder
+WORKDIR /tmp/facets-preview
+RUN mkdir -p inst/facets_qc \
+    && cp -r facets_qc/. inst/facets_qc/ \
+    && R CMD INSTALL .
+
+FROM runtime_base AS final
+
+COPY --from=htslib_builder /usr/lib/libhts.so* /usr/lib/
+COPY --from=htstools_builder "/tmp/htstools-snp_pileup_${HTSTOOLS_VERSION}/snp-pileup" /usr/bin/
+COPY --from=htstools_builder "/tmp/htstools-snp_pileup_${HTSTOOLS_VERSION}/ppflag-fixer" /usr/bin/
+COPY --from=facets_suite_builder "/tmp/facets-suite-${FACETS_SUITE_VERSION}/" /usr/bin/facets-suite/
+COPY --from=facets_preview_builder /tmp/facets-preview/ /usr/bin/facets-preview/
+COPY --from=facets_preview_builder /usr/local/lib/R/site-library/ /usr/local/lib/R/site-library/
+
+SHELL ["/bin/bash", "-o", "pipefail", "-c"]
+
+WORKDIR /
+
+RUN rm -rf /tmp/* /var/tmp/* \
+    && ldconfig \
+    # ---------- verify R packages ----------
+    && R --vanilla -e " \
+    pkgs <- list( \
+        plyr          = NULL, \
+        diptest       = NULL, \
+        DT            = NULL, \
+        reshape2      = NULL, \
+        rtracklayer   = NULL, \
+        data.table    = Sys.getenv('DATA_TABLE_VERSION'), \
+        pctGCdata     = Sys.getenv('PCTGCDATA'), \
+        facets        = Sys.getenv('FACETS_VERSION'), \
+        facets2n      = NULL, \
+        facetsSuite   = Sys.getenv('FACETS_SUITE_VERSION'), \
+        facetsPreview = NULL \
+    ); \
+    failed <- c(); \
+    for (p in names(pkgs)) { \
+        if (!requireNamespace(p, quietly = TRUE)) { \
+            failed <- c(failed, p); \
+        } else { \
+            ver <- as.character(packageVersion(p)); \
+            expected <- pkgs[[p]]; \
+            cat(sprintf('  %-20s %s\n', p, ver)); \
+            if (!is.null(expected) && ver != expected) \
+                stop(sprintf('Version mismatch for %s: expected %s, got %s', p, expected, ver)); \
+        } \
+    }; \
+    if (length(failed) > 0) \
+        stop('The following R packages failed to load: ', paste(failed, collapse = ', ')); \
+    message('All R package version checks passed.')" \
+    # ---------- verify the pinned facets2n flat-genome dipLogR guard ----------
+    && R --vanilla -e " \
+    src <- paste(deparse(facets2n:::findDiploidLogR), collapse = ' '); \
+    if (!grepl('length(cn2logR) == 0', src, fixed = TRUE)) \
+        stop('facets2n is missing the flat-genome dipLogR guard; wrong revision installed.'); \
+    message('facets2n flat-genome dipLogR guard present.')" \
+    # ---------- verify binary tools ----------
+    && snp-pileup --help 2>&1 | grep -q 'vcf file' \
+    && ppflag-fixer --help 2>&1 | grep -q 'input file' \
+    && /usr/bin/facets-suite/run-facets-wrapper.R --help | grep -q 'COUNTS_FILE' \
+    # 2N-specific flags: assert this is facets-suite-2n and not stock facets-suite
+    && /usr/bin/facets-suite/run-facets-wrapper.R --help | grep -q 'facets2n-lib-path' \
+    && /usr/bin/facets-suite/run-facets-wrapper.R --help | grep -q 'MandUnormal' \
+    && /usr/bin/facets-suite/snp-pileup-wrapper.R --help | grep -q 'unmatched-normal-BAMS'
+
+CMD ["/bin/bash"]