Filter protein

What is Filter protein?

Protein sequences acquired from databases, alignment outputs, or manual entry frequently contain characters that fall outside the expected amino acid alphabet. Digits from copy-pasted annotations, whitespace from text editors, stop codon asterisks, and ambiguity codes can all cause downstream tools to reject input or produce incorrect results. Filter Protein strips or replaces these characters based on a configurable allowed set, producing clean sequences ready for analysis.

Amino acid alphabets

Not all 26 letters correspond to amino acids. The standard genetic code encodes 20 amino acids, each with a one-letter designation originally proposed by Margaret Oakley Dayhoff:

A (Ala), C (Cys), D (Asp), E (Glu), F (Phe), G (Gly), H (His), I (Ile), K (Lys), L (Leu), M (Met), N (Asn), P (Pro), Q (Gln), R (Arg), S (Ser), T (Thr), V (Val), W (Trp), Y (Tyr).

Beyond these, IUPAC nomenclature reserves additional letters for special or ambiguous residues:

Code	Meaning
B	Aspartate or asparagine (Asx)
Z	Glutamate or glutamine (Glx)
J	Leucine or isoleucine (Xle)
U	Selenocysteine (Sec)
O	Pyrrolysine (Pyl)
X	Unknown or non-standard residue

The distinction matters for filtering. A sequence from mass spectrometry where D and N cannot be distinguished will use B; a sequence from a structure database might include U for selenocysteine. Choosing the wrong filter mode discards valid information.

How to use Filter Protein online

Filter Protein runs entirely in the browser on ProteinIQ. No data leaves the client, making it suitable for proprietary or sensitive sequences. Results appear instantly.

Input

Paste one or more protein sequences in FASTA format or plain text, or upload a file. The tool deliberately skips sequence validation on input, since cleaning invalid characters is the purpose. Header lines are kept exactly as written, and lines starting with # or ; are skipped with a warning. FASTQ, GenBank, and UniProt or EMBL flat files are rejected with a message naming the tool that extracts their sequences first.

Format	Extensions
Text	.txt
FASTA	.fasta, .fa, .fas

Maximum file size: 10 MB.

Filter modes

Each mode defines a set of characters to keep. Everything outside the set is either deleted or replaced.

Mode	Characters kept	Typical use
Standard 20 amino acids	A, C, D, E, F, G, H, I, K, L, M, N, P, Q, R, S, T, V, W, Y	Preparing input for most analysis tools
Standard 20 + stop codon (*)	Standard 20 plus *	Preserving stop codons from translation output
IUPAC amino acid codes	Standard 20 plus B, J, O, U, X, Z	Retaining ambiguity codes and rare amino acids
IUPAC + gap characters	IUPAC set plus - and .	Cleaning aligned sequences without removing gaps
All letters (A-Z)	Any letter	Removing only non-alphabetic noise
All letters + gap	Any letter plus - and .	Minimal filtering of aligned data
Remove whitespace only	Everything except spaces, tabs, newlines	Reformatting pasted text
Remove digits only	Everything except 0-9	Stripping line numbers or position annotations
Remove digits and whitespace	Everything except digits and whitespace	Combined cleanup
Custom allowed characters	User-specified character set	Any non-standard filtering requirement

Replacement options

When a character falls outside the allowed set, it can be handled in two ways: deletion (the character is removed entirely, shortening the sequence) or replacement (the character is substituted with a placeholder). Spaces, line breaks, digits, and invisible characters from copied documents are formatting, so they are always deleted and never replaced. The three Remove ... modes only delete, so Replace invalid with is hidden for them.

Setting value	Effect
Delete	Remove the character, reducing sequence length
X or x	Replace with X (conventional for unknown residue)
N or n	Replace with N
- or .	Replace with a gap character
?	Replace with question mark
*	Replace with stop codon symbol
Custom character	Replace with any user-specified character

Replacing with X rather than deleting preserves positional information. This matters when filtered sequences need to remain aligned with other data, such as secondary structure annotations or conservation scores that are indexed by residue position.

Output formatting

Setting	Default	Description
Output case	Uppercase	Case of the kept residues; replacement characters are inserted as chosen
Preserve FASTA headers	Enabled	Retain original > header lines in the output
Line length	80	Characters per line; set to 0 to output each sequence on a single line

The result opens on Preview with a one-line summary of sequences, output length, characters removed, and characters replaced. Files holds the filtered FASTA (a .txt file when headers are turned off) and run.log, which lists the settings, removal counts for each character, and input and output length per record.

Warnings appear above the preview when residues were deleted (which shifts later positions), when a record had no sequence or nothing left after filtering, or when comment lines were skipped. If nothing is left at all, the run stops with an error.

Applications

Sequence filtering is a routine preprocessing step in bioinformatics workflows. A few common scenarios:

• Input sanitization: Tools like AlphaFold 2, ESMFold, and ProteinMPNN require strictly valid amino acid sequences and will fail on unexpected characters.
• Post-translation cleanup: DNA to Protein translation may produce stop codon symbols (*) that need removal before structural prediction.
• Alignment post-processing: Multiple sequence alignment tools such as MUSCLE5, MAFFT, and Clustal Omega insert gap characters that must be stripped before submitting to other analyses.
• Database submission: Sequence databases typically accept only the standard 20 amino acid letters and reject files containing formatting artifacts.
• Quality control: The character counts in run.log show exactly which non-standard characters were present and how many, useful for auditing data provenance.