The cellranger pipeline outputs an HDF5 file containing per-molecule information for all molecules that contain a valid barcode, a valid UMI, and were assigned with high confidence to a gene or Feature Barcode. This HDF5 file contains data corresponding to the observed molecules, as well as data about the libraries and feature set(s) used (general information about the HDF5 file format available here). This file is called molecule_info.h5 in cellranger count and sample_molecule_info.h5 in cellranger multi outputs.

Copy
(root)
    ├─ barcode_idx
    ├─ barcode_info	[HDF5 group]
    │   ├─ genomes
    │   └─ pass_filter
    ├─ barcodes
    ├─ count
    ├─ feature_idx
    ├─ features	[HDF5 group]
    │   ├─ _all_tag_keys
    │   ├─ target_sets [for Targeted Gene Expression or Flex]
    │   │    └─ [target set name]
    │   ├─ feature_type
    │   ├─ genome
    │   ├─ id
    │   ├─ name
    │   ├─ pattern [Feature Barcode only]
    │   ├─ read [Feature Barcode only]
    │   └─ sequence [Feature Barcode only]
    ├─ gem_group
    ├─ library_idx
    ├─ library_info
    ├─ metrics_json [HDF5 dataset; see below]
    ├── probe_idx          ---------------------|
    ├── probes [HDF5 group]                     |
    │   ├── feature_id                          | [For Flex, Cell Ranger v7.1+]
    │   ├── feature_name                        |
    │   ├── probe_id                            |
    │   └── region         ---------------------|
    ├─ umi
    └─ umi_type

You can examine the contents of the H5 file using software such as HDFView or the h5dump command, as demonstrated below to show the file contents of the entire H5 object:

Copy
h5dump -n molecule_info.h5

    HDF5 "molecule_info.h5" {
    FILE_CONTENTS {
    group      /
    dataset    /barcode_idx
    group      /barcode_info
    dataset    /barcode_info/genomes
    dataset    /barcode_info/pass_filter
    dataset    /barcodes
    dataset    /count
    dataset    /feature_idx
    group      /features
    dataset    /features/_all_tag_keys
    dataset    /features/feature_type
    dataset    /features/genome
    dataset    /features/id
    dataset    /features/name
    dataset    /gem_group
    dataset    /library_idx
    dataset    /library_info
    dataset    /metrics_json
    dataset    /umi
    dataset    /umi_type
    }
    }

The following HDF5 datasets in the molecule info file correspond to columns of a table. Each row of that table corresponds to a unique (UMI, cell-barcode, feature) tuple indicating the feature best supported by the reads (i.e., including PCR duplicates) assigned to that UMI and cell-barcode.

In addition, the molecule info file has datasets corresponding to information about the libraries, barcodes, and feature set(s) that were used in the analysis, as described below.

Experiment reference

The barcodes, library_info, and metrics_json datasets contain information about the experiments contained in this analysis:

Observed cell-barcodes

The HDF5 group barcode_info contains information regarding the barcodes that were called as cells during the analysis. This HDF5 group contains:

Feature reference

The HDF5 group features contains information regarding the feature reference(s) used for the analysis. The datasets within the features group represent columns in a table containing one row per feature. Values in the feature_idx column described in the previous section provide indices into the rows of this table of features.

In addition to the columns described below, user-specified tags may also be present. The dataset _all_tag_keys contains a list of user-specified tags as well as built-in tags (genome, pattern, read, and/or sequence).

Probe reference

Present for Flex analysis with Cell Ranger v7.1 and later. The HDF5 group probes contains information regarding the probe set used for the analysis. The datasets within the probes group represent the columns in a table containing one row per probe. Values in the probe_idx column described in the previous section provide indices for the rows of this table of probes.

2-bit encoding

The UMI sequences are 2-bit encoded as follows:

• Each pair of bits encodes a nucleotide (0="A", 1="C", 2="G", 3="T").
• The least significant byte (LSB) contains the 3'-most nucleotides.

Note that the cell-barcode sequences do not have this encoding. Instead, they are stored as plain strings in the library_info HDF5 group.