Skip to content

Abundance

The ts.abun.est_from_fasta and ts.abun.est_from_gtf modules estimate the relative abundance of each transcript in order to proportionally sample molecules from a Fasta/GTF file. The total number of molecules is controlled by --molecule-count. It makes “one cell” abundance (the true molecule count setting for this sample).

  • Sampling methods
    • dirichlet
    • lognormal
    • zipf
    • uniform
  • Selecting methods
    • random (uniform)
    • first (first occurrence)

Estimation from Fasta

Abundance can be estimated using a transcript fasta file through the following function.

Download the fasta file.

wget -O gencode.v48.transcripts.fa.gz https://ftp.ebi.ac.uk/pub/databases/gencode/Gencode_human/release_48/gencode.v48.transcripts.fa.gz

If you want to generate 50 transcripts from the file and the total expression value is 10000, do

import tresor as ts

res = ts.abun.est_from_fasta(
    fasta_fpn="D:/Document/Programming/Python/umiche/umiche/data/r1/cm/gencode.v48.transcripts.fa.gz",
    num_selected_mols=50,
    select_mode="random",
    num_total_mols=10000,
    use_pipe_if_present=True,
    pipe_index=0,
    strip_version=True,
    dist="dirichlet",
    alpha=1.0,
    min_count=0,
    sort_output_by="count",
    output_tsv="D:/Document/Programming/Python/umiche/umiche/data/r1/cm/tksm/abundance.tsv",
    subset_ids_txt="D:/Document/Programming/Python/umiche/umiche/data/r1/cm/tksm/refs.sub.ids.fa.txt",
    seed=1,
)
print("Top5:")
for i in range(min(5, len(res['ids']))):
    print(res['ids'][i], res['counts'][i])

The format of abundance.tsv file: tab-delimited, must contain at least Transcript ID and Abundance as two columns.

abundance.tsv looks like

transcript  count
ENST00000769524 935
ENST00000816532 780
ENST00000820931 767
...
ENST00000754605 6
ENST00000672107 4

Estimation from GTF

Download a GTF file from Genecode.

wget -O gencode.v48.annotation.gtf.gz https://ftp.ebi.ac.uk/pub/databases/gencode/Gencode_human/release_48/gencode.v48.annotation.gtf.gz

# wget -O gencode.v48.basic.annotation.gtf.gz https://ftp.ebi.ac.uk/pub/databases/gencode/Gencode_human/release_48/gencode.v48.basic.annotation.gtf.gz

import tresor as ts

res = ts.abun.est_from_gtf(
    gtf_fpn="D:/Document/Programming/Python/umiche/umiche/data/r1/cm/tksm/gencode.v48.annotation.gtf",
    num_selected_mols=5,
    select_mode="random",
    num_total_mols=10,
    strip_version=True,
    dist="dirichlet",
    alpha=1.0,
    min_count=0,
    sort_output_by="count",
    include_noncoding=False,
    output_tsv="D:/Document/Programming/Python/umiche/umiche/data/r1/cm/tksm/abundance.pure.tsv",
    versioned_out="D:/Document/Programming/Python/umiche/umiche/data/r1/cm/tksm/abundance.versioned.tsv",
    subset_ids_txt="D:/Document/Programming/Python/umiche/umiche/data/r1/cm/tksm/refs.sub.ids.gtf.txt",
    seed=1,
)
print("Top5:")
for i in range(min(5, len(res['ids']))):
    print(res['ids'][i], res['counts'][i])

abundance.pure.tsv looks like

ENST00000477535 4
ENST00000304081 3
ENST00000509522 2
ENST00000593404 1
ENST00000610462 0

abundance.versioned.tsv looks like

ENST00000593404.5   878
ENST00000620405.1   732
ENST00000530097.1   720
...
ENST00000542002.5   317
ENST00000519650.5   286

TPM

Simple TPM normalization

\[ TPM_i = \frac{count_i}{\sum count} \times 10^6 \]

Explanation

  • \(TPM_i\): TPM value of the i-th transcript
  • \(count_i\): Read count of the i-th transcript
  • \(Σ(count)\): Sum of counts across all transcripts
  • \(1e6\): Scaling factor (per million)

You can add a cell barcode (e.g., ATGCATGCATGCATGC) to the sampled transcripts in cell_barcode to denote from which cells they are derived.

import tresor as ts

res = ts.abun.to_tpm(
    counts_tsv="D:/Document/Programming/Python/umiche/umiche/data/r1/cm/tksm/abundance.versioned.tsv",
    out_tsv="D:/Document/Programming/Python/umiche/umiche/data/r1/cm/tksm/abundance.versioned.tpm.tsv",
    cell_barcode="ATGCATGCATGCATGC",
    include_header=False,
    float_format="{:.6f}"
)

abundance.versioned.tpm.tsv looks like

ENST00000477535.5   400000.000000   ATGCATGCATGCATGC
ENST00000304081.9   300000.000000   ATGCATGCATGCATGC
ENST00000509522.6   200000.000000   ATGCATGCATGCATGC
ENST00000593404.5   100000.000000   ATGCATGCATGCATGC
ENST00000610462.1   0.000000    ATGCATGCATGCATGC