06 Bioinformatics Fundamentals

Bioinformatics and biological data

Bioinformatics uses computational methods to store, organize, compare, and interpret biological data. Sequence analysis is a core example: researchers examine DNA, RNA, or protein sequences to identify similarities, predict features, and develop hypotheses about biological function. Results depend on the data and settings used, so analyses should be reproducible and their limitations understood.

Sequence data and identifiers

Biological data include nucleotide sequences from DNA and RNA, protein sequences, sequencing reads, genome assemblies, sequence annotations, and measurements such as gene-expression levels. A sequence is represented as a string of symbols: DNA commonly uses A, C, G, and T; RNA uses A, C, G, and U; proteins use one-letter amino-acid codes. A symbol such as N may represent an unknown nucleotide.

Sequence records can include annotations as well as sequence data. Annotations describe features such as gene locations, coding regions, or organism names. An accession number identifies a database record, and a versioned accession can help specify the sequence revision that was analyzed.

FASTA and FASTQ formats

FASTA stores a sequence with a description line beginning with > followed by the sequence. It is useful for reference sequences and many analysis tools.

FASTQ stores sequencing reads together with a per-base quality score. It is commonly used for raw sequencing data and, unlike FASTA, retains information about how confident the sequencing instrument was in each base call.

Choosing a sequence database

Biological sequence databases differ in the data they collect and how records are produced. GenBank is an archival collection of submitted nucleotide sequences. RefSeq provides a non-redundant set of reference sequences—including genomes, transcripts, and proteins—with varying levels of annotation and curation.

SRA (Sequence Read Archive) stores high-throughput sequencing data and associated information about samples and experiments. UniProt provides protein sequences and functional information.

Choose a database according to the question being asked. Raw-read archives are useful for finding sequencing experiments; reference collections are often helpful for comparisons against representative sequences; and protein resources provide protein-focused annotations. Check the organism, record description, accession and version, evidence, and database update or release information.

Preparing and comparing sequences

A basic sequence-analysis workflow proceeds from inspecting the input to choosing a comparison and evaluating the output in context.

1. Validate and inspect sequences. Check that symbols and format are appropriate, record sequence lengths, and examine read quality when available. Sequencing reads may need quality trimming or other preprocessing before downstream analysis.

2. Transform sequences when needed. DNA may be reverse-complemented to compare the opposite strand. A coding DNA sequence can be translated into amino acids using a genetic code; this can help compare a nucleotide sequence with protein records.

3. Align sequences. An alignment places residues in corresponding positions and may insert gaps to represent possible insertions or deletions. A global alignment compares sequences across their full lengths and is useful for similar, same-length sequences. A local alignment finds the most similar matching regions within sequences of different lengths.

4. Search a database. BLAST compares a query sequence with sequences in a selected database and reports regions of local similarity. Choose the program to match the query and database, such as nucleotide-to-nucleotide or protein-to-protein.

Interpreting alignment and search results

Alignment scores depend on the scoring system, including how matches, mismatches, and gaps are valued. BLAST results include an E-value: the number of alignments with a score at least as strong that would be expected by chance in a search of that size. A smaller E-value generally indicates a less chance-like match.

Interpret results using more than the E-value. Consider alignment length, percent identity, query coverage, database choice, and biological context. Similarity to an annotated sequence can suggest a function, but does not by itself prove that the query has the same function.

For example, a short DNA query may match one region of a much longer gene. A strong local alignment can indicate a related segment without showing that the complete sequences—or their functions—are equivalent. Compare the result with additional records and evidence before drawing conclusions.

Reproducibility and interpretation

Bioinformatics applies computational and statistical approaches to biological information. DNA, RNA, proteins, sequencing reads, and annotations are stored in formats and databases suited to different purposes. Introductory sequence analysis involves checking data, applying appropriate transformations, aligning sequences, and searching databases. Results provide evidence for biological interpretation, not automatic proof, and should be reported with enough information to make the analysis reproducible.