Python Bio Variables

Variables and Data Types

When to Use

• Writing or reviewing introductory Python bioinformatics scripts that assign sequences, counts, or scores to variables.
• Explaining why alias = original doesn't copy a list/dict, or why seq = seq + "AAA" doesn't mutate the original string.
• Choosing between int, float, str, bool, None for genomic data (coordinates, GC content, E-values, sequences, missing annotations).
• Slicing/indexing DNA/RNA/protein strings (start/stop codons, reverse complement setup, coding regions).
• Parsing simple text formats (FASTA headers, delimited fields) with split()/join()/strip().

Version Compatibility

Python ≥3.9 (f-strings, arbitrary-precision int — no special version dependency; syntax here works on any Python 3.x).

Prerequisites

None — this is foundational Python, no external packages required. Useful next step: bio-sequence-manipulation-seq-objects for Bio.Seq objects once past raw strings.

Naming rules and conventions

Rules (breaking these is a SyntaxError):

• Names may contain letters, digits, and underscores; must start with a letter or underscore (not a digit).
• Names are case-sensitive (gene and Gene are different variables).
• Python keywords (if, for, class, ...) cannot be used as names.

Conventions:

• snake_case for variables/functions: gene_name, sequence_length.
• UPPER_CASE for constants: AVOGADRO = 6.022e23.
• Descriptive names: gc_content, not gc or x.

# Good vs. bad variable names
sequence_length = 1500        # good: descriptive, snake_case
gene_name = "TP53"
melting_temperature = 65.5

x = 1500                      # bad: meaningless
seqLen = 1500                  # bad: camelCase, not snake_case
three = 1                      # bad: misleading name for the value 1

# 2nd_gene = "EGFR"           # SyntaxError: cannot start with a digit

Data Types Overview

def describe_types(values):
    """Print the runtime type of each value in an iterable.

    Mirrors a common bioinformatics debugging step: confirming a parsed
    field (from a VCF, FASTA header, or config file) came back as the
    expected type rather than a string.
    """
    for v in values:
        print(f"{str(v):25s} -> {type(v).__name__}")

describe_types([15_000_000, 0.52, "Escherichia coli", True, None])
# 15000000                  -> int
# 0.52                      -> float
# Escherichia coli          -> str
# True                      -> bool
# None                      -> NoneType

Variables are references, not boxes

Goal: understand why aliasing a list/dict silently shares mutations, while aliasing a string/number does not. Approach: inspect id() before and after reassignment vs. in-place mutation.

def show_reference_semantics():
    """Demonstrate immutable reassignment vs. mutable in-place mutation."""
    # Strings are immutable: concatenation creates a NEW object
    sequence = "ATGCGATCG"
    before_id = id(sequence)
    sequence = sequence + "AAA"          # new string, sequence now points elsewhere
    assert id(sequence) != before_id

    # Lists are mutable: alias shares the SAME object
    reads = ["read1", "read2"]
    alias = reads                        # alias points to the same list
    alias.append("read3")
    assert reads == ["read1", "read2", "read3"]  # mutation visible via original name

    # Multiple-assignment gotcha: a = b = [] shares ONE list, not two
    coverage_a = coverage_b = []
    coverage_a.append(10)
    assert coverage_b == [10]            # surprising if you expected independence

show_reference_semantics()

Numeric types for bioinformatics

# int: unlimited precision, used for lengths, counts, coordinates
sequence_length = 3_088_286_401   # human genome length in bp
read_depth = 30                   # sequencing coverage (30x)
print(f"Genome: {sequence_length:,} bp, target coverage: {read_depth}x")

# float: measurements, scores, probabilities
gc_content = 0.508                # GC fraction
e_value = 1.5e-42                 # BLAST E-value, scientific notation
print(f"GC%: {gc_content * 100:.1f}%  E-value: {e_value:.2e}")

# Floating-point precision: 0.1 + 0.2 != 0.3 exactly (binary representation)
result, expected, tolerance = 0.1 + 0.2, 0.3, 1e-9
is_close_enough = abs(result - expected) < tolerance   # always compare with a tolerance
assert is_close_enough

String slicing and parsing (sequences, FASTA headers)

Goal: extract start/stop codons, reverse a sequence, and parse a FASTA header. Approach: string[start:stop:step] — start inclusive, stop exclusive, negative step reverses.

def parse_orf(dna):
    """Extract start codon, stop codon, and coding region from a raw ORF string.

    dna: full open reading frame including start (ATG) and a stop codon.
    Returns a dict with start_codon, stop_codon, coding_region, reversed.
    """
    return {
        "start_codon": dna[0:3],
        "stop_codon": dna[-3:],
        "coding_region": dna[3:-3],
        "has_valid_stop": dna.endswith(("TAA", "TAG", "TGA")),
        "reversed": dna[::-1],
    }

orf = parse_orf("ATGAAACCCGGGTAA")
assert orf["start_codon"] == "ATG"
assert orf["stop_codon"] == "TAA"
assert orf["coding_region"] == "AAACCCGGG"
assert orf["has_valid_stop"] is True

def parse_fasta_header(header):
    """Parse a UniProt-style FASTA header: '>db|accession|entry_name description'."""
    parts = header.lstrip(">").split("|")
    db, accession, rest = parts[0], parts[1], parts[2]
    entry_name = rest.split()[0]
    return {"database": db, "accession": accession, "entry_name": entry_name}

result = parse_fasta_header(">sp|P04637|P53_HUMAN Cellular tumor antigen p53")
assert result == {"database": "sp", "accession": "P04637", "entry_name": "P53_HUMAN"}

# DNA -> RNA transcription via replace(); strip() for cleaning file lines
dna = "ATGCGATCG"
rna = dna.replace("T", "U")
assert rna == "AUGCGAUCG"
clean_line = "  ATGCGATCG  \n".strip()
assert clean_line == "ATGCGATCG"

Pitfalls

• Aliasing shares mutable objects: alias = original does not copy a list/dict — both names reference the same object; mutating through one is visible through the other. Immutable types (str, int, float, tuple) don't have this problem.
• a = b = [] creates ONE shared list, not two independent empty lists. Use a, b = [], [] for independence.
• Mutable default arguments: never write def f(x=[]) — the default list is created once and shared across calls. Use def f(x=None) and set x = x or [] inside.
• Off-by-one errors: Python slicing is half-open [start, stop), but genomic coordinates (GFF, 1-based VCF POS) are often 1-based inclusive — convert explicitly (gff_start - 1) before slicing a Python string.
• Deep vs. shallow copy: list.copy() only copies the top level; nested structures (list of lists, dict of lists) need copy.deepcopy().
• Floating-point equality: never compare floats with ==; use abs(a - b) < tolerance.

See Also

• bio-sequence-manipulation-seq-objects — move from raw strings to Bio.Seq objects.
• bio-sequence-manipulation-sequence-slicing — deeper slicing patterns for sequences.
• bio-sequence-manipulation-reverse-complement — building on string reversal shown here.
• bio-sequence-io-read-sequences — reading real FASTA/FASTQ files instead of inline strings.