Sequences are sent to our server only to run the analysis, processed in memory, and never stored or logged.
CCCTTCTACGTAACCTAAAATTACGCTTGCAACGCAGGTGAAAC
Batch analysis
Analyze multiple sequences at once and compare them in a table.
CCCTTCTACGTAACCTAAAATTACGCTTGCAACGCAGGTGAAAC
Pairwise alignment
Compare two sequences to find similarities and differences.
CCCTTCTACGTAACCTAAAATTACGCTTGCAACGCAGGTGAAAC
Advanced primer design
Finds the best forward and reverse primer pairs for PCR.
CCCTTCTACGTAACCTAAAATTACGCTTGCAACGCAGGTGAAAC
BLAST
Paste a DNA, RNA or protein sequence. Aaron's Archive detects which one it is and runs the matching search (blastn or blastp) for you. Requires an internet connection.
CCCTTCTACGTAACCTAAAATTACGCTTGCAACGCAGGTGAAAC
Phylogenetic tree
Paste 2–100 sequences in FASTA format to build an evolutionary tree using UPGMA clustering on distances from an alignment of the full sequences. You get sequence relationships, branch lengths and an exportable Newick string.
CCCTTCTACGTAACCTAAAATTACGCTTGCAACGCAGGTGAAAC
Mutation effect simulator
See what a DNA change does to the protein.
GGTGAT
CRISPR guide RNA finder
Lists candidate guide RNAs for the region you want to edit.
GACTTA
Guide
Getting started
What is Aaron's Archive?
Aaron's Archive is a bioinformatics tool that reads DNA, RNA, and protein sequences and calculates everything about them — composition, structure, genes, melting temperature, primers, and more. Think of it like a super-powered calculator designed specifically for genetic data.
You paste a DNA sequence → Aaron's Archive tells you its GC%, where genes are, what proteins it would make, where restriction enzymes cut it, and much more — all instantly.
What is a DNA sequence?
DNA (Deoxyribonucleic Acid) is the molecule that carries all genetic instructions. It's made of 4 bases: A (Adenine), T (Thymine), G (Guanine), C (Cytosine). A DNA sequence is a string of these letters. In this tool, you just type them in.
Example: ATGAAAGCAATTTTCGTACT... (this would be part of a real gene)
What is RNA? How is it different from DNA?
RNA (Ribonucleic Acid) is DNA's messenger. When a gene is "read," DNA is copied into RNA (called transcription). RNA uses U (Uracil) instead of T. So AUGCAUGC is RNA, ATGCATGC is DNA. Aaron's Archive auto-detects which one you pasted.
Central dogma: DNA → RNA → Protein. Aaron's Archive can analyze all three steps.
What is a protein sequence?
Proteins are made of 20 amino acids, each with a single-letter code: A C D E F G H I K L M N P Q R S T V W Y. The order (sequence) of amino acids determines the protein's shape and function. Aaron's Archive detects protein sequences automatically.
MTEYKLVVVGAGG... is the start of the KRAS protein (mutated in many cancers)
What is FASTA format?
FASTA is the universal format for biological sequences. It starts with a "header" line beginning with > (the name), then the sequence on the next line. Aaron's Archive accepts plain sequences AND FASTA format — it automatically strips the header.
>Insulin_gene ATGGCCCTGTGGATGCGCCTCCTG...
Mutation simulator & CRISPR
What do silent, missense and nonsense mean?
Silent: the DNA changes but the amino acid stays the same (the genetic code has spare codons). Missense: one amino acid is swapped for another. Nonsense: the change creates a stop signal, so the protein ends early. A frameshift adds or removes bases that are not a multiple of 3, so every codon after it is read wrongly.
KRAS c.35G>A turns GGT (glycine) into GAT (aspartate): p.Gly12Asp, a famous cancer mutation.
How do I write a change like c.35G>A?
The c. means "coding DNA". The number counts from the A of the ATG start codon (base 1). 35G>A says base 35 is a G and becomes an A. Other forms: c.35del (delete), c.35_36insGG (insert between 35 and 36), c.35dup (duplicate), c.35_37delinsAA (replace).
If the base you name does not match your sequence, the tool tells you what is really there.
What are Grantham and BLOSUM62?
Two ways to score how different two amino acids are. Grantham (0 to 215) measures chemical difference: bigger means a more drastic swap. BLOSUM62 counts how often evolution accepts the swap: positive means common, negative means rare. They hint at impact but cannot prove it.
What is a CRISPR guide RNA and a PAM?
Cas9 is guided to DNA by a short RNA (the guide, about 20 letters) that matches the target. Cas9 only cuts if the target is followed by a short motif called the PAM (NGG for SpCas9). So guides can only be designed where the PAM exists. The tool scans both strands for every valid site.
Cut position: SpCas9 cuts about 3 bp before the PAM.
How is the guide score worked out?
It starts at 100 and loses points for GC content outside 40 to 70%, TTTT (stops U6 transcription), GGGG, long single-base runs, guides that can fold on themselves, and exact or near copies elsewhere in your sequence. It is a design checklist, not a prediction of cutting efficiency, and it cannot see off-target sites in the rest of the genome.
DNA/RNA analysis terms
GC Content (%)
The % of bases that are G or C (vs A or T). G and C pair with 3 hydrogen bonds (stronger), while A and T pair with 2 (weaker). Higher GC% = more stable, harder to melt apart. Human DNA averages ~41% GC.
If a 100 bp sequence has 25G + 20C, GC% = 45%. A GC-rich sequence needs higher PCR temperatures.
Melting Temperature (Tm)
The temperature at which half the DNA double helix separates into single strands. Higher GC% = higher Tm. Three estimates are shown: the nearest-neighbor model (SantaLucia 1998, the standard for 8–60 nt primers, here at 50 mM Na⁺ and 250 nM oligo, so values are comparable with Primer3 / IDT-style calculators), a GC-content formula for longer DNA, and the Wallace rule (only valid under 14 nt). Critical for designing PCR conditions.
A primer with 50% GC typically melts around 55–62°C. PCR annealing temperature = Tm − 5°C
ORFs — Open Reading Frames
A continuous stretch of DNA that could be translated into a protein. It starts with ATG (start codon = Methionine) and ends with TAA/TAG/TGA (stop codons). DNA has 6 possible reading frames (+1,+2,+3 on top strand, -1,-2,-3 on bottom). Longer ORFs are more likely to be real genes.
Frame +1 reads: ATG|AAA|GCA|... in triplets starting at position 1. Frame +2 starts at position 2, etc.
Restriction Enzyme Sites
Restriction enzymes are "molecular scissors" — bacterial proteins that cut DNA at specific recognition sequences. EcoRI cuts at GAATTC. Scientists use these to cut DNA at predictable locations for cloning (inserting genes into plasmids). The map shows where each enzyme would cut your sequence.
If you want to clone a gene into a vector, you need restriction sites that: (1) are present in your gene region, (2) are also in the vector's multiple cloning site, and (3) don't cut inside your gene of interest.
CpG Islands
Regions where CG dinucleotides appear more frequently than expected. Found near most gene promoters (start sites) in vertebrates. CpG methylation (adding a chemical group to C) silences genes — this is called epigenetics. Roughly 60–70% of human gene promoters lie in a CpG island, so an island is a useful hint of a promoter (methylating its cytosines usually switches the gene off). Islands are defined by composition alone, not by function.
If your sequence has a CpG island, a gene promoter may be nearby, but confirm it with other evidence.
GC Skew
GC Skew = (G − C) / (G + C). Measures whether a region has more G or C. In bacterial genomes, GC skew changes sign at the replication origin and terminus — making it useful to find where DNA replication starts. Positive = more G, Negative = more C.
GC skew = +0.2 means 60% G vs 40% C in that window
Microsatellites / SSRs
Short sequences of 1–6 bases repeated many times in a row: like CACACACACACA (CA × 6), or a mononucleotide run like AAAAAAAAAA. Called Simple Sequence Repeats (SSRs). They are highly variable between people — different people have different numbers of repeats. Used in DNA fingerprinting, paternity testing, forensics, and population genetics.
The forensic TV show "DNA match" uses STR (Short Tandem Repeat = microsatellite) analysis. Two people sharing the same repeat pattern across 20 loci = virtually certain match.
Shannon Entropy / Complexity %
Measures how "random" or "mixed" the sequence is. Maximum complexity = all 4 bases appear equally (like random DNA). Low complexity = repetitive (AAAAAA or ATATATATAT). Measured in bits. The % compares your sequence to the theoretical maximum. Repetitive regions are often non-coding.
AAAAAAAAAA → 0%. ATATATATAT → 50%. ATGCATGCATGC → 100%, because the score only counts how often each base occurs and ignores their order, so a perfectly repeating pattern can still score 100%. Random genomic DNA → about 95–100%.
Codon Usage
A codon is 3 DNA letters that code for one amino acid. There are 64 codons but only 20 amino acids, so most amino acids have multiple codons (synonyms). Different organisms prefer different synonymous codons — called codon bias. If you're producing a human protein in bacteria, "rare codons" slow down protein production. Codon optimization replaces rare codons with preferred ones.
Leucine can be coded by: TTA, TTG, CTT, CTC, CTA, CTG. Humans prefer CTG most. E. coli also prefers CTG but dislikes CTA.
Promoter Elements (TATA box, Kozak, etc.)
Specific DNA sequences that control gene expression. The TATA box (TATAAA) sits ~30 bp before gene start where RNA polymerase binds. The Kozak sequence surrounds the ATG start codon and affects translation efficiency. Shine-Dalgarno is the ribosome binding site in bacteria (AGGAGG ~10 bp upstream).
A TATA-like match is only a hint. A pattern this short also turns up by chance about twice per 10,000 bases of random DNA, so a hit upstream of an ORF supports, but does not prove, that it is an expressed gene.
Single-stranded RNA folds back on itself because A pairs with U and G pairs with C (and, more weakly, G with U) within the same molecule. This creates stem-loop (hairpin) structures. Aaron's Archive uses ViennaRNA, which finds the most stable fold, the minimum free energy (Zuker's algorithm with Turner thermodynamic parameters), and reports it in kcal/mol: more negative means more stable. If ViennaRNA is not available on the server, a simple Nussinov method that only maximises the number of base pairs is used instead, and the page says which was used. The structure is shown in dot-bracket notation: ( and ) are paired, . is unpaired. mRNA structure affects translation, tRNA structure allows it to carry amino acids, ribozymes use RNA structure for catalysis.
( ( ( . . . ) ) ) means the first 3 and last 3 bases pair with each other, forming a stem with 3 unpaired bases in the loop
Protein analysis terms
Isoelectric Point (pI)
The pH at which the protein has zero net charge. Below pI → protein is positive. Above pI → protein is negative. Critical for: purification by ion-exchange chromatography (at pI the protein won't bind), isoelectric focusing gel, predicting protein behavior at physiological pH 7.4.
pI 5 = acidic protein (negative at body pH 7.4). pI 9 = basic protein (positive at body pH). At pI, proteins often precipitate.
GRAVY Score (Hydrophobicity)
Grand Average of Hydropathicity. Positive = the protein prefers fat/membrane environments (hydrophobic). Negative = the protein prefers water (hydrophilic). Uses the Kyte-Doolittle scale for each amino acid.
Rough guide: most soluble proteins score between about -1 and 0, and integral membrane proteins usually score above 0 (bacteriorhodopsin: +0.71); a whole-protein GRAVY above +1 is very rare. Transmembrane helices are not judged from this whole-protein number but from a 19-residue window averaging at least +1.6.
Instability Index
An estimate of the protein's stability in a test tube (Guruprasad 1990, as in ExPASy ProtParam), calculated from which pairs of neighbouring amino acids occur. Below 40 = predicted stable, above 40 = predicted unstable. It is a statistical prediction from sequence, not a measurement, and it says nothing about how fast a protein is broken down inside a cell.
Use it as a rough warning when planning protein production: a high value suggests the protein may need extra care (low temperature, protease inhibitors, a stabilising tag).
Aromaticity
The fraction of aromatic amino acids (Phenylalanine F, Tyrosine Y, Tryptophan W). These absorb UV light at 280 nm — this is how we measure protein concentration in lab (spectrophotometry with a NanoDrop machine). High aromaticity also stabilizes protein structure through π-stacking interactions.
W absorbs most strongly, then Y. To measure protein concentration without a standard curve (1 cm pathlength): concentration (mg/mL) = A280 ÷ [(5500 × W + 1490 × Y + 125 × cystines) ÷ molecular weight]
Extinction Coefficient (ε280)
How strongly a protein absorbs UV light at 280nm per mole, in M⁻¹cm⁻¹ — used with the Beer-Lambert law (A=εcl) to calculate protein concentration directly from an absorbance reading, without needing a standard curve. Calculated from Trp, Tyr, and cystine (disulfide-bonded Cys) content using the Gill & von Hippel / Pace method — the same one ExPASy ProtParam uses.
Concentration (M) = A280 ÷ ε280. If A280 = 0.5 and ε280 = 37,000, concentration ≈ 13.5 μM.
Aliphatic Index
A measure of thermostability based on the relative volume occupied by aliphatic side chains (Ala, Val, Ile, Leu) — proteins with a higher aliphatic index tend to remain folded and stable across a wider temperature range. Introduced by Ikai (1980).
Thermostable/thermophilic proteins often score 80-100+. Typical mesophilic (room-temperature) proteins score around 60-90.
Secondary Structure (Chou-Fasman)
Proteins fold into characteristic shapes. α-Helix = spring/coil (like a spiral staircase). β-Sheet = flat strands side by side (like a folded paper fan). Chou-Fasman propensities score each amino acid's tendency to sit in helices vs. sheets, from statistics of known structures. The numbers here are whole-sequence averages, so they describe a composition tendency, not a predicted structure.
Alanine (A) strongly prefers helix. Valine (V) and Isoleucine (I) prefer beta-sheet. Glycine (G) and Proline (P) break helices.
Transmembrane Helices
Proteins embedded in cell membranes cross the lipid bilayer as α-helices made of hydrophobic amino acids. Found with the Kyte-Doolittle rule (1982): any 19-residue stretch whose average hydropathy is at least 1.6 is flagged, and overlapping stretches are merged. GPCRs (like adrenergic receptors and opsins) have 7 TM helices; ion channels and transporters have other numbers. This simple rule misses some helices, so confirm with TMHMM or DeepTMHMM.
Seven candidates would fit a GPCR-like receptor and one would fit a single-pass membrane protein, but a missed or extra helix is common with this rule, so treat the count as a hint.
Signal Peptide
A short stretch (15–30 aa) at the protein's N-terminus (beginning) that acts as a "postal address" telling the cell to send this protein to the ER for secretion or membrane insertion. The signal peptide is cleaved off after delivery. A simple rule-based check can only hint that one is present; confirm with SignalP.
Insulin has a signal peptide (first 24 aa). After cleavage: preproinsulin → proinsulin → insulin. Antibodies (IgG) also have signal peptides — they're secreted proteins.
Charge vs pH curve
Shows how the total charge of the entire protein changes as pH goes from 0 to 14. At low pH (acidic), the protein is positively charged. At high pH (basic), it's negative. Where the curve crosses zero = the pI. The slope at pH 7.4 tells you how sensitive the protein is to pH changes. Used to optimize protein purification conditions.
This curve is used in the lab to choose the right pH buffer for protein storage and purification.
Hydrophobicity Profile chart
A sliding window (9 aa at a time) plot of hydrophobicity along the protein sequence, for visualizing the general shape of the hydrophobic/hydrophilic pattern. The peaks shown here are a guide for the eye — the actual transmembrane-helix calls (listed in their own table) use a wider 19-residue window whose average must reach +1.6. Valleys below zero = hydrophilic surface-exposed regions. Helps identify functional domains, binding sites, and membrane topology.
For a membrane receptor: you'd see peaks (TM helices) alternating with valleys (hydrophilic loops connecting them).
Tools explained
Primers — what are they?
Primers are short DNA sequences (18–25 bases) that act as starting points for DNA copying in PCR. You need two: a forward primer (matches the top strand) and a reverse primer (matches the bottom strand). The region between them gets amplified (copied millions of times). Good primers need: GC% 40-60%, Tm 55–65°C, no self-folding (hairpins), no pairing with each other (dimers).
PCR is how COVID tests amplify viral DNA, how forensic DNA is tested, and how researchers clone genes. Primers define exactly what gets amplified.
Primer Hairpin & Dimer problems
A hairpin forms when part of the primer folds back on itself and self-pairs — this means the primer is "busy" folding instead of binding to your template, reducing efficiency. A self-dimer means two copies of the same primer pair together. A heterodimer means the forward and reverse primers bind each other instead of the template. All three reduce or eliminate PCR product.
A primer ending in GCTCGC that starts with GCGAGC would form a strong hairpin — those sequences are reverse complements of each other.
Sequence Alignment explained
Alignment lines up two sequences to find similarities. It can insert gaps (-) to make the sequences match better. It's like a "diff" tool for DNA — showing what's identical, what mutated, and what was inserted/deleted. Used to find evolutionary relationships, conserved functional sites, and mutations between strains or species.
Aligning the TP53 gene from a tumor vs. normal tissue shows the exact mutation (e.g., G→A at position 248).
Smith-Waterman (Local alignment)
Finds the best matching REGION between two sequences, ignoring non-matching ends. Best when you think sequences are only similar in part (e.g., comparing a protein domain to a full genome). Returns the single best local match. This algorithm underlies BLAST.
Use local alignment when: comparing a 50 aa conserved domain to a 500 aa protein, or finding a gene fragment within a genome.
Needleman-Wunsch (Global alignment)
Aligns the ENTIRE length of both sequences end-to-end, penalizing every gap. Best when comparing sequences of similar length that are related throughout (same gene from two species). Shows total evolutionary divergence across the full sequence.
Use global alignment when: comparing the full insulin gene from human vs. pig, or checking how similar two strains of the same virus are across their whole genome.
BLAST — Basic Local Alignment Search Tool
BLAST searches NCBI's enormous database (billions of known sequences from millions of organisms) to find what's most similar to your sequence. It's like a Google search for DNA. Results tell you: what gene/protein this matches, what organism it's from, how similar it is (E-value), and where the alignment is.
You paste an unknown sequence → BLAST says: "99% identical to human TP53 tumor suppressor, chromosome 17p13.1." Mystery solved!
BLAST E-value (most important field)
Expect value = how many hits this good you'd expect by random chance. Lower = more significant. E=0 or <1e-100 = perfect/near-perfect, definitely not random. E=1e-10 to 1e-50 = very significant. E=0.01 to 1 = possibly significant. E>10 = likely random. When comparing against databases of billions of sequences, even E=0.001 can be meaningful.
Finding E=3.2e-98 for a 200bp sequence means this match is astronomically unlikely by chance — it's definitely a real biological homology.
Ramachandran Plot — what it shows
A plot of every residue's two backbone twist angles — φ (phi, rotation around N–Cα) and ψ (psi, rotation around Cα–C) — which together determine how the protein chain folds at that point. Not every angle combination is physically possible without atoms colliding, so the plot is shaded into favored (ideal, low-strain), allowed (still reasonable), and unshaded outlier regions. This is the same idea structural biologists use (e.g. via MolProbity) to judge whether a solved or predicted structure is trustworthy — the shaded regions are contours of the MolProbity Top8000 (2018) reference data (2-degree grid, one set of contours per residue class), drawn from a pre-computed grid, so the split is a good guide but not a replacement for running MolProbity itself.
>90% of residues in the favored region is typical of a good-quality structure. A high outlier percentage can mean a genuinely flexible loop, an active-site distortion — or a lower-resolution/lower-confidence structure worth double-checking.
3D Structure Viewer (AlphaFold & PDB)
After BLAST identifies a gene, Aaron's Archive can pull up its predicted or experimentally solved 3D structure — AlphaFold for AI-predicted models, RCSB PDB for real crystal/cryo-EM structures. AlphaFold structures are colored by pLDDT confidence: blue/high-confidence regions are trustworthy, orange/red low-confidence regions (often flexible loops or disordered regions) are less certain.
A protein with mostly blue coloring and a compact fold is a confident, likely-accurate prediction. Long red stretches often mean intrinsically disordered regions that don't have one fixed shape in reality.
Batch FASTA analysis
Instead of analyzing one sequence at a time, paste multiple sequences in FASTA format and get a comparison table. Useful for: comparing GC content across many genes, screening multiple variants for restriction sites, comparing sequences from different samples in an experiment.
Analyzing 20 different bacterial isolates for antibiotic resistance gene content — one click gives you a comparison table.
BLOSUM62 matrix (for protein alignment)
A scoring table used for protein alignment that scores each possible amino acid substitution based on how often it occurs in evolution. Conservative substitutions (similar amino acids like Leucine↔Isoleucine) score high (+2). Radical substitutions (very different like Tryptophan→Glycine) score negative (-2). Calibrated on proteins sharing ~62% identity.
Leucine to Isoleucine (both hydrophobic, similar size): score +2. Alanine to Tryptophan (tiny→huge, hydrophobic→aromatic): score -3.
Lab context for students
PCR — Polymerase Chain Reaction
A technique to copy a specific DNA region millions of times. Uses DNA polymerase and your two primers. The 3 temperature cycles: (1) 94–98°C — Denaturation: melts DNA apart. (2) 50–65°C — Annealing: primers bind their matching sites. (3) 72°C — Extension: polymerase builds new strand. After 30 cycles = 2³⁰ copies = ~1 billion copies from one original. Used in COVID tests, forensics, gene cloning, sequencing prep.
Why GC% matters for PCR
GC-rich primers need higher annealing temperatures (they're "stickier"). AT-rich primers melt too easily at high temperatures. If forward and reverse primer Tm differ by >5°C, one primer won't bind efficiently at the other's optimal temperature → weak or no PCR product. That's why ΔTm is shown in the primer results.
Restriction cloning — why it matters
To insert a gene into a plasmid for protein expression: (1) Cut gene and plasmid with the same restriction enzyme (creates matching "sticky ends"). (2) Mix and let sticky ends hybridize. (3) DNA ligase seals the joins. The restriction map helps you find which enzymes cut your gene in useful places (and which would destroy it).
Molecular Weight in biology
Mass of a molecule. The commonly cited "1 kbp DNA ≈ 660 kDa" rule is for double-stranded DNA (~650-660 Da per base pair, i.e. per two nucleotides). This tool reports the mass of your sequence as a single strand (~310-330 Da per nucleotide, using each base's actual monoisotopic-adjacent average weight) — the convention used by oligo-synthesis calculators (e.g. IDT's OligoAnalyzer) for a pasted sequence. For the double-stranded mass of the same region, roughly double this figure. For proteins: 1 average amino acid ≈ 110 Da, so a 100 aa protein ≈ 11 kDa. Used when running gel electrophoresis (size determines how far molecules migrate), and calculating how much of a compound to add in experiments (molarity).