Sequence Slicing - Usage Guide

Overview

This skill enables AI agents to help you extract, slice, and concatenate biological sequences with Biopython. It covers bare-Seq slicing, carrying a sub-region of an annotated record (with quality scores and features) into a new record, coordinate-system conversion, and joining sequences. The central judgment it encodes is what metadata survives a SeqRecord slice and how to avoid the 0-based vs 1-based off-by-one.

Prerequisites

pip install biopython

Quick Start

Tell your AI agent what you want to do:

• "Extract positions 100-200 from this sequence"
• "Pull out this CDS but keep its per-base quality scores"
• "The gene is at 1234-5678 in the GFF; extract it with the right coordinates"
• "Splice these exon regions into one transcript"
• "Join these sequences with an NNN linker"
• "Split this sequence into 100 bp chunks"

Example Prompts

Basic Extraction

"Extract nucleotides 100 to 500 from my sequence and return a Seq."

Coordinates From a File

"My GFF says the feature spans 1234 to 5678 (1-based inclusive). Slice the matching subsequence without an off-by-one."

"This interval came from a BED file (0-based). Extract it correctly."

Keep Annotations and Qualities

"Slice positions 100-400 out of this FASTQ record but keep the per-base quality scores."

"Extract this region of a GenBank record and carry the organism and taxonomy into the new record."

Splicing and Features

"Splice these three exon coordinate pairs into a single mRNA sequence."

"Extract every CDS feature from this GenBank file, strand-aware."

Splitting and Joining

"Split this coding sequence into codons."

"Concatenate these sequences with NNNNNN between each."

"Give me 50 bp on each side of position 1000."

What the Agent Will Do

1. Decide whether the request needs a bare Seq slice or a SeqRecord slice (qualities/features to keep).
2. Convert coordinates to 0-based half-open, subtracting 1 from the start for 1-based GFF/GenBank input.
3. Apply Python slicing, or feature.extract() for strand-aware/joined locations.
4. Copy the annotations dict explicitly when carrying a sub-region into a new record.
5. Return the extracted, spliced, or concatenated sequence or record.

Coordinate System Notes

Python and Biopython slice 0-based and half-open: seq[0] is the first base, seq[0:3] returns three bases (0, 1, 2), seq[100:200] returns 100 bases. File formats differ:

• GenBank / EMBL / GFF / GTF / VCF feature coordinates are 1-based and inclusive. Convert by subtracting 1 from the START only: seq[start-1:end]. The exclusive Python end already cancels the inclusive file end, so the end needs no change.
• BED coordinates are already 0-based half-open and slice directly.
• Bio.SeqFeature locations store 0-based start and Python-style end, so feature.extract(record.seq) slices correctly and handles strand and joined exons.

What a SeqRecord Slice Keeps and Drops

record[start:end] preserves id, name, description, and molecule_type, and auto-slices per-letter data (PHRED quality in letter_annotations) to match. It keeps only features fully contained in the range, with recalculated locations. It SILENTLY drops the annotations dict (organism, taxonomy, references), dbxrefs, and any feature straddling the boundary. A stride such as record[::2] drops features entirely. Copy what you need: sub.annotations = record.annotations.copy().

Tips

• Always clarify whether coordinates are 0-based or 1-based before slicing; the off-by-one is silent.
• Slice the SeqRecord (not record.seq) when per-base qualities or contained features must ride along.
• Copy annotations after a SeqRecord slice; slicing always drops it without warning.
• seq[::-1] reverses but does NOT complement; use reverse_complement() for the opposite strand.
• Use feature.extract(record.seq) for any strand-aware or joined/compound location instead of a manual slice.
• sum() over slices needs a Seq('') start value; the default 0 cannot be added to a Seq.
• Guard str(record.seq) with len() first; undefined content raises UndefinedSequenceError.

Related Skills

• seq-objects - Create Seq/SeqRecord objects and handle undefined sequence content
• reverse-complement - Reverse-complement an extracted region
• transcription-translation - Translate an extracted CDS or spliced transcript
• sequence-io/read-sequences - Parse GenBank/FASTQ records to slice
• genome-intervals/gtf-gff-handling - Read 1-based GFF/GTF feature coordinates before slicing
• alignment-files/sam-bam-basics - Extract sequences from BAM regions with samtools