# A Beginner's Guide to Using Expasy: Tools and Databases By the end of this guide you will be able to take a [protein sequence](/blog/guides/protein-sequence), run it through the core ExPASy bioinformatics tools, and read the results with confidence. You will estimate molecular weight and isoelectric point with ProtParam, find domains with PROSITE, predict protease cleavage sites with PeptideCutter, build a 3D model with SWISS-MODEL, and browse the Swiss-Prot knowledgebase behind it all. Everything runs in a browser. You need no programming experience, no Linux machine, and no software installation. Here is what to have on hand before you start: - A modern web browser - A protein sequence in FASTA format (the first worked example below gives you one you can copy) - A text editor such as Notepad, TextEdit (in plain text mode), or VS Code for holding sequences - Roughly one hour for the full walkthrough This article serves as a beginner's guide to using ExPASy as a first stop when you learn bioinformatics, and it doubles as a practical companion for anyone taking an introductory [bioinformatics course](/blog/careers/bioinformatics-courses-and-certifications-a-practical-guide) that uses web-based protein analysis. ## What Is ExPASy and Why Beginners Start Here ExPASy stands for the Expert Protein Analysis System. It is a web server maintained by the SIB Swiss Institute of Bioinformatics, and it provides access to a large set of databases and analytical tools dedicated to proteins and proteomics [1]. The site has a long history. It started operating in 1993 as the first World Wide Web server in the life sciences, which means many working biologists today learned sequence analysis on it [1]. Three things make ExPASy a good starting point when you learn bioinformatics. First, the databases are curated by people, not just scripts. The Swiss-Prot protein knowledgebase provides manually annotated entries for all species, with concentrated effort on model organisms so that representative members of every protein family carry high quality annotation [2]. Swiss-Prot later merged with TrEMBL and PIR to form UniProt, the single centralized resource for protein sequences and functional information that most of the field now uses [2]. Second, the tools and databases are tightly interlinked. The proteomics tools were designed to read the annotations in Swiss-Prot, so a prediction made by an ExPASy tool can use known facts about the protein family rather than working from raw sequence alone [1]. Third, ExPASy is proven teaching material. Undergraduate courses have built entire inquiry-based laboratories around these tools, sending students to NCBI, ExPASy, the [Protein Data Bank](/blog/guides/protein-data-bank), ClustalW and KEGG to investigate disease-associated proteins and defend hypotheses about mutations [3]. If a tool can carry a first-year lab course, it can carry your first solo analysis. The site also grows with the field. Since 2016 there has been a dedicated Glycomics@ExPASy tab that centralizes web-based glycoinformatics resources, designed to be friendly to both glycoscientists and protein scientists [4], and the broader Glyco@Expasy initiative connects databases and tools that highlight interactions between glycoproteins and carbohydrate-binding proteins [5]. You do not need any of that on day one, but it is worth knowing the site extends beyond amino acid sequences. ## Before You Start: Get a Sequence Every ExPASy protein tool takes an [amino acid sequence](/blog/guides/amino-acid-sequence) as its primary input. Sequences are exchanged in FASTA format, a plain text format where the first line starts with a greater-than sign and holds a description, and every following line holds the sequence itself, written with the single-letter amino acid codes. Here is a real example you can use for the whole walkthrough. This is human ubiquitin, a small, extremely well-studied protein of 76 amino acids: ``` >Human ubiquitin, 76 amino acids MQIFVKTLTGKTITLEVEPSDTIENVKAKIQDKEGIPPDQQRLIFAGKQLEDGRTLSDYNIQKESTLHLVLRLRGG ``` To get your own sequence, go to UniProt (uniprot.org), type a protein name such as "human ubiquitin" or "human lysozyme" into the search box, open an entry, and copy the sequence. Prefer entries reviewed by curators, which UniProt marks as reviewed and which correspond to Swiss-Prot quality annotation [2]. A note on formatting, because it causes the most beginner errors. The FASTA header line is optional for most ExPASy tools, but the sequence characters are not. Valid characters are the 20 standard amino acid letters (ACDEFGHIKLMNPQRSTVWY), plus ambiguous codes such as X, B, Z, and U for selenocysteine. Spaces, numbers, dashes, and paste artifacts from PDFs will trigger errors. ## ExPASy Bioinformatics Tools Step by Step: A First Protein Analysis This walkthrough follows the sequence a working analyst would use on an unfamiliar protein. It mirrors the order used in published characterizations, where authors typically start with physicochemical properties, move to domains and motifs, then to structure, and finally to function [6]. ### Step 1: Open ProtParam and Paste the Sequence ![Calculation showing 76 residues times 110 Da giving about 8,400 Da, close to the ProtParam value of 8564.84 Da for ubiquitin.](/article-images/illustrations/expasy-bioinformatics-tools-protparam-mass.webp) ProtParam computes physicochemical properties that can be derived directly from the amino acid sequence. Go to expasy.org, find ProtParam under the protein analysis tools, paste your FASTA text into the box, and submit. For the ubiquitin sequence above, the top of the output looks like this: ``` Number of amino acids: 76 Molecular weight: 8564.84 Theoretical pI: (computed value for your sequence) Total number of negatively charged residues (Asp + Glu): (computed value) Total number of positively charged residues (Arg + Lys): (computed value) Extinction coefficient: (computed under three Cys assumptions) Instability index: (computed value) Aliphatic index: (computed value) Grand average of hydropathicity (GRAVY): (computed value) ``` The molecular weight of 8564.84 Da for this exact 76-residue sequence is a textbook value, and it is the fastest sanity check that you pasted the right protein. A useful mental model: the average amino acid residue contributes roughly 110 Da after the loss of water during peptide bond formation, so a 76-residue protein should land near 8,400 Da. ProtParam computes the exact value from the actual residue composition, which is why it reads slightly higher than the back-of-envelope estimate. Here is how to read each field. - **Molecular weight** tells you what the unmodified polypeptide chain weighs. Compare it against the expected size of your protein of interest on an SDS-PAGE gel. - **Theoretical pI** is the pH at which the protein carries no net charge. A protein is positively charged below its pI and negatively charged above it. This drives choices in ion exchange chromatography and buffer design. - **Charged residue counts** summarize Asp plus Glu versus Arg plus Lys. A protein with more negatively charged residues than positively charged ones, as one published SARS-CoV protein characterization reported, hints at how the protein will behave in charge-based separations [6]. - **Extinction coefficient** estimates absorbance at 280 nm, which is how you measure protein concentration with a spectrophotometer. The tool reports values under three assumptions about cysteine oxidation, so read the one that matches what you know about disulfide bonds in your protein. - **Instability index** is a sequence-derived estimate of stability. Values above 40 are conventionally read as a hint toward instability in the test tube. This is not trivia. In one published study of chlorogenic acid biosynthesis enzymes cloned from Taraxacum mongolicum, the enzyme with the lowest measured catalytic activity was the one with the higher instability index of 45.49, and the authors tied that weaker activity to the predicted instability [7]. Treat the index as a warning light, not a verdict. - **Aliphatic index** reflects the relative volume occupied by aliphatic side chains and is widely used as a positive indicator of thermostability. A high aliphatic index was cited as one reason a characterized SARS-CoV protein was considered stable [6]. - **GRAVY**, the grand average of hydropathicity, sums the Kyte-Doolittle hydropathy values and divides by the length. Negative values indicate hydrophilic proteins. One published analysis of a parasite serine protease inhibitor combined ExPASy predictions with expression work and reported the protein as a stable hydrophilic protein, consistent with a negative GRAVY reading [8]. None of these numbers are measurements. They are sequence-derived calculations, and the next steps put them in context. ### Step 2: Probe Hydropathy Along the Sequence With ProtScale ProtParam gives you one hydropathy number for the whole protein. ProtScale gives you the whole curve. Choose the Kyte-Doolittle scale, pick a sliding window size, and the tool plots hydropathy at each position. Short windows around 9 residues highlight local hydrophobic patches, which can suggest membrane-spanning regions or buried cores. Longer windows smooth the curve and show overall tendencies. Reading the plot is simple. Regions that stay above the hydrophobic threshold across a stretch of roughly 20 residues are candidates for transmembrane helices, since a membrane crossing requires about that many residues to span the bilayer. A soluble protein like ubiquitin shows no such stretch. A receptor or channel will typically show several. This step matters because it explains contradictory ProtParam numbers. A membrane protein can carry a positive or near-zero GRAVY even though most of its surface is hydrophilic, because the transmembrane segment dominates the average. ### Step 3: Scan for Motifs and Domains With ScanProsite PROSITE is one of the original ExPASy databases, a collection of biologically meaningful patterns and profiles tied to protein families, domains, and functional sites [1]. ScanProsite searches your sequence against them. Paste the ubiquitin sequence and submit. The tool reports any matching patterns. The classic beginner example is the N-glycosylation sequon, written in PROSITE syntax as: ``` N-{P}-[ST]-{P} ``` Reading the syntax: an N, then any residue except proline, then serine or threonine, then any residue except proline. Any asparagine fitting that rule is a potential N-glycosylation site, although only a subset of candidate sites is actually glycosylated in the cell. Many published analyses explicitly enumerate such predicted glycosylation and phosphorylation sites alongside domain content when characterizing protein families [9]. For a protein you know nothing about, the PROSITE hits are often the first functional clues. A match to a kinase domain profile, a zinc finger, or a serine protease signature tells you what experiments to plan next. This is exactly the pattern that published characterizations of unannotated proteins follow, using sequence-derived domain and motif information to argue for a probable function before any experiment [6]. ### Step 4: Predict Where Proteases Will Cut With PeptideCutter PeptideCutter reports potential cleavage sites for a large panel of proteases and chemicals. The tool sees real use in peptide research. In one published pipeline for identifying antimicrobial peptides from Camelus dromedarius, researchers used PeptideCutter on ExPASy alongside BLASTP, conserved domain searches, and [signal peptide](/knowledge/molecular-biology/signal-peptide) prediction to shortlist candidate cathelicidin peptides, then synthesized the top candidates and validated their antibacterial effects experimentally [10]. The clearest way to learn the tool is a worked example with trypsin. Trypsin cleaves on the C-terminal side of lysine (K) and arginine (R), unless the next residue is proline (P). Take this 12-residue test sequence: ``` MAGKPRDLAAKR ``` Walk the rule by hand: - K at position 4 is followed by P, so no cut there - R at position 6 is followed by D, so cut after position 6 - K at position 11 is followed by R, so cut after position 11 - R at position 12 ends the sequence Submit the sequence to PeptideCutter with trypsin selected. The expected fragments are: ``` MAGKPR (residues 1 to 6) DAAK (residues 7 to 11) R (residue 12) ``` If the tool returns exactly these fragments, you understand both the rule and the tool. This exercise also builds intuition for mass spectrometry, because trypsin digestion before LC-MS analysis produces precisely these kinds of peptides, and in-silico digestion lets you predict the peptide map before the run. ### Step 5: Check for a Signal Peptide and Topology Before modeling a structure, ask where the protein ends up. Signal peptides at the N-terminus direct secretion and are cleaved off in the mature protein, which means the full precursor sequence and the mature chain differ in both length and computed properties. ExPASy provides topology and signal prediction resources among its protein analysis tools [1], and published pipelines routinely run such predictions alongside physicochemical analysis before committing to expression experiments [10], [8]. The practical consequence for you as a beginner: when a published recombinant protein study reports the length and molecular weight of the construct they expressed, check whether they included or excluded the signal peptide. That single decision changes every downstream number, and mismatched constructs are a classic source of bands on a gel that do not match predictions. ### Step 6: Build a 3D Model With SWISS-MODEL SWISS-MODEL is a fully automated homology modeling server, and the SWISS-MODEL repository was already part of the ExPASy database collection in the original description of the server [1]. Homology modeling rests on a simple idea: proteins with similar sequences fold similarly, so if your protein has a relative in the Protein Data Bank with a solved structure, you can borrow that structure and adjust it. The workflow: 1. Go to SWISS-MODEL from the ExPASy structure tools. 2. Paste the same target sequence, or paste a UniProt accession and let the server fetch the canonical sequence. 3. Submit the Automated Mode job. The server searches for structural templates, builds models against the best ones, and returns coordinates you can download and view. 4. Read the model quality estimates in the results page. Higher confidence means the model is built on a close template and covers more of your sequence. Where coverage is partial, the server is telling you that only part of your protein resembles anything of known structure, and the unmodeled regions should not be interpreted structurally. This is a standard component of published protein characterization pipelines. The Taraxacum study used SWISS-MODEL alongside ProtParam to describe its biosynthetic enzymes before cloning and expressing them [7]. The Strongyloides protease inhibitor study used ExPASy, SWISS-MODEL and other tools to predict structure and antigenic epitopes, then cloned the gene into a pET28a [expression vector](/knowledge/molecular-biology/expression-vector) and confirmed the recombinant protein by Western blotting and mass spectrometry [8]. The hypothetical SARS-CoV protein study modeled secondary and tertiary structure with several tools and then assessed the models with dedicated quality evaluators such as PROCHECK and ProSA-web before discussing function [6]. Notice the pattern in all three: modeling came before experiments, and experiments had the final word. ### Step 7: Translate a DNA Sequence With the Translate Tool When your starting material is a [nucleotide sequence](/knowledge/molecular-biology/nucleotide-sequence) rather than a protein, the Translate tool converts it to the amino acid sequence using the standard genetic code. Paste a coding sequence: ``` ATGGCTGGCAACTTTGAC ``` The output shows six reading frames. The one that begins with M and runs without stop codons to the end is the open reading frame, and it yields: ``` M A G N F D ``` In practice you will paste hundreds or thousands of bases, and the tool will show you all six frames side by side. Frames containing frequent stop codons are almost certainly not the coding frame. If your nucleotide sequence came from a eukaryotic gene with introns, translation will produce nonsense, and you should first fetch the coding sequence (CDS) rather than genomic DNA. ### Step 8: Search for Similar Sequences When you meet an unfamiliar sequence, similarity search is the single most informative first move, and similarity searches were among the core analysis tasks ExPASy was built around [1]. Run a search against the protein databases with your sequence. The result ranks known proteins from most to least similar, with aligned regions and statistical scores. The top hits tell you the likely identity, and the pattern of hits across a family tells you how conserved your protein is. One caution ties back to the databases. Hits against Swiss-Prot, with its manual annotation, are far more trustworthy as functional leads than hits against automatically translated entries [2]. When two hits conflict, the reviewed entry wins. ## A Quick Reference Table of Core ExPASy Tools | Tool | What it computes | Input you need | How to read the result | |, - |, - |, - |, - | | ProtParam | Molecular weight, pI, instability index, aliphatic index, GRAVY, extinction coefficient | A raw protein sequence or UniProt accession | Sequence-derived estimates. Use as sanity checks and experiment planners, not measurements | | ProtScale | Sliding-window hydropathy or other scales along the sequence | A protein sequence, a scale, a window size | Sustained hydrophobic stretches suggest transmembrane regions | | ScanProsite | Matches to PROSITE patterns and profiles | A protein sequence | Each hit names a family, domain or site. Verify the hit makes biological sense for your protein | | PeptideCutter | Cleavage sites for many proteases | A protein or peptide sequence | Fragments under each selected enzyme. Compare against your in-silico digestion expectations | | Similarity search (BLAST and related) | Ranked database hits for your sequence | A sequence | Top reviewed hits suggest identity and family | | SWISS-MODEL | A homology-based 3D model with quality estimates | A sequence or accession | Trust the modeled regions per the quality metrics. Unmodeled regions are unknown, not invisible | | Translate | All six reading frames of a DNA sequence | A nucleotide sequence | The open reading frame runs from M to a stop codon without internal stops | | Glycomics@ExPASy | A hub of glycoinformatics databases and tools | Varies by resource | Start here when glycans enter your question [4] | ## The Databases Behind the Tools Tools compute. Databases know. The strength of ExPASy comes from the pairing, and four databases deserve a beginner's attention. **Swiss-Prot and the road to UniProt.** Swiss-Prot entries are manually annotated by curators who read the literature, reconcile contradictions, and attach experimental evidence to specific claims. The annotation effort concentrates on model organisms so that every protein family has at least some high quality representative members, and direct links to specialized databases connect sequences with current knowledge in their field [2]. For plant scientists this took the form of a dedicated program annotating proteins from Arabidopsis thaliana, with the explicit hope that the accumulated Arabidopsis knowledge and the ExPASy software tools would help identify functions in other plants [2]. In 2002 through 2003 the Swiss-Prot, TrEMBL and PIR resources merged into UniProt, which became the single authoritative protein sequence and function resource [2]. When you search UniProt today and see a "reviewed" flag, you are looking at Swiss-Prot quality annotation. **PROSITE.** A curated library of sequence patterns and profiles for domains and functional sites, integrated into ExPASy from the start [1]. It is the fastest way to attach a name to an unannotated region. **ENZYME.** An enzyme nomenclature database linking proteins to their official classification and catalyzed reactions [1]. When your BLAST hit is an enzyme, ENZYME tells you what reaction the family performs. **SWISS-2DPAGE and the SWISS-MODEL repository.** These support the proteomics and structure sides of the server, mapping proteins onto two-dimensional gel data and providing precomputed models, respectively [1]. The glycoinformatics corner shows how the site has kept expanding. Glycomics@ExPASy was created in 2016 to centralize web-based glycoinformatics resources developed by an international network of glycoscientists, hosting mostly SIB-maintained databases and tools alongside reference resources popular in the glycomics community [4]. The Glyco@Expasy collection deliberately emphasizes interactions between glycoproteins such as cell surface receptors and carbohydrate-binding proteins, and its published demonstrations walk through querying and visualizing the N-glycome of prostate specific antigen and the O-glycome of human serum proteins [5]. The design philosophy is explicit: make the tools as interactive as possible, based on database connectivity, so researchers can explore data and build hypotheses rather than fetch static pages [4]. ## ExPASy Bioinformatics Tools Tips From Real Workflows ![Decision path for a new protein: unknown sequence to similarity search, DNA to Translate, structure questions to SWISS-MODEL, otherwise ProtParam.](/article-images/illustrations/expasy-bioinformatics-tools-first-tool-choice.webp) The most useful tips come from studying how experienced analysts chain these tools. The following habits show up repeatedly in published characterizations. **Tip 1: Fix the sequence before you fix the interpretation.** Confirm the protein identity with a similarity search before you write a single interpretation sentence about it. The camel antimicrobial peptide study began with BLASTP and conserved domain searches before any property prediction [10]. Identity first, properties second. **Tip 2: Combine tools in a fixed pipeline and reuse it.** Published papers converge on a recognizable sequence of steps: physicochemical properties with ProtParam, secondary structure with a predictor such as SOPMA or PSIPRED, domains with CD Search or PROSITE, tertiary structure with SWISS-MODEL or HHpred, then model quality assessment with tools such as PROCHECK and ProSA-web [6]. Adopting a fixed pipeline makes your work reproducible and makes your notebooks comparable across proteins. One human MEF2 [transcription factor](/knowledge/molecular-biology/transcription-factor) analysis used ExPASy together with CDD and SABLE, then built a phylogenetic tree from the alignments, finding conserved MADS domains across all family members and abundant predicted phosphorylation sites [9]. The pipeline, not any single tool, produced the insight. **Tip 3: Use the accession, not just the sequence, whenever possible.** Tools that accept a UniProtKB accession pull in annotation along with sequence, so the analysis inherits the curated knowledge that Swiss-Prot provides [2]. Raw pasted sequences compute the same numbers but carry no identity. **Tip 4: Remember that predictions precede experiments, and experiments decide.** This is the strongest tip on this page. Every study cited in this guide that used ExPASy tools went on to test predictions at the bench, whether by cloning genes into pET30a vectors and measuring enzymatic product formation [7], expressing a recombinant protein in E. coli BL21 and verifying it by Western blot and mass spectrometry [8], or synthesizing predicted antimicrobial peptides and testing them with colony-forming assays, electron microscopy and hemolysis assays [10]. In silico analysis generates hypotheses. It does not close them. **Tip 5: Treat the instability index as a design input.** If ProtParam flags a candidate protein as unstable, that prediction can be weighed when choosing among paralogs for expression. The Taraxacum study did exactly this kind of reasoning when it connected one enzyme's weaker activity to its higher instability index of 45.49 [7]. The number is a heuristic, but heuristics are useful when you must pick one clone from three. **Tip 6: Match the prediction set to the question.** Antigenic epitope prediction mattered for the parasite vaccine-oriented study [8]. Extinction coefficient mattered for anyone measuring concentration spectrophotometrically. Cleavage prediction mattered for peptide design [10]. Ask what number would actually change your next decision, then find the tool that produces it. **Tip 7: For teaching and self-study, work a real disease protein end to end.** The undergraduate course that inspired generations of similar labs gave students a cDNA sequence encoding a mutant human protein, sent them through NCBI and ExPASy and the Protein Data Bank and ClustalW and KEGG, and had them present a hypothesis for the mutation's effect [3]. That exercise remains one of the best ways to learn bioinformatics, and you can run it on yourself with any protein from a paper you are reading. ## Common Mistakes and Limitations Every beginner hits the same handful of walls. Here they are, with fixes. **Mistake 1: Pasting a nucleotide sequence into a protein tool.** ProtParam expects amino acid letters. Feed it DNA and you will get an error rather than a translation, because the tool does not guess your intent. The fix is to run the sequence through the Translate tool first, identify the correct reading frame, and paste the resulting protein sequence. **Mistake 2: Dirty FASTA text.** Copying sequences from PDFs, Word documents, or web pages imports line breaks, spaces, page numbers, and non-ASCII characters. The tool rejects the submission or silently computes on a truncated sequence. Fix: keep sequences in plain text files, one sequence per file, and check the reported sequence length against your expectation before reading any output. A 76-residue protein should report exactly 76 residues. **Mistake 3: Analyzing the precursor when you mean the mature chain, or the reverse.** A secreted protein's full sequence includes a signal peptide that the cell removes. ProtParam on the precursor gives a molecular weight and pI that will never match the purified mature protein. Fix: look up the mature chain boundaries in the UniProt entry annotation and analyze the region you actually have in hand. **Mistake 4: Reading every prediction as a fact.** The instability index, aliphatic index, GRAVY, and secondary structure predictions are all correlations derived from sequences of known behavior. They carry no information about your protein's specific post-translational modifications, proteolytic processing, oligomerization, or binding partners, because none of that is in the sequence string. The extinction coefficient offers a concrete example of why this matters. It depends on the number of tryptophan, tyrosine, and cysteine residues, and the tool reports three values under three cysteine oxidation states. Choosing the right one requires knowledge the tool does not have. **Mistake 5: Over-trusting structural models.** A SWISS-MODEL output is only as good as its template. Regions of your protein that have no structural relative come back unmodeled, and quality estimates exist precisely so you do not quote a model as if it were an X-ray structure. Published work treats model quality assessment as its own step, using dedicated evaluators after modeling [6]. Do the same, and never build a mechanistic claim on an unmodeled loop. **Mistake 6: Confusing a motif hit with a function.** An N-glycosylation sequon in a sequence is a candidate site, not a proven glycosylation. PROSITE pattern matches are hints that become claims only with supporting evidence from the entry annotation or experiment. The curated Swiss-Prot annotation exists to help you make this distinction, which is part of why manually annotated entries carry more weight than automatically generated ones [2]. **Mistake 7: Forgetting that the endpoint is a hypothesis.** The most consequential limitation of any prediction tool is that the final answer comes from the lab. The peptide candidates predicted in the camel study earned their conclusions only after circular dichroism spectroscopy, bacterial killing assays, and membrane disruption imaging [10]. The enzyme candidates predicted in the dandelion study earned theirs only after in vitro reactions with caffeoyl-CoA and quinic acid measured actual product formation [7]. Plan the experiment at the same time as the prediction, and interpret borderline predictions conservatively. **Mistake 8: Working outside your competence.** Sequence predictions can inform experimental design, but interpreting biological or clinical meaning from them requires appropriate training and, for applied work, a qualified supervisor or specialist. If a result would drive a decision about health, disease, or treatment, route it through a professional rather than a web tool. A final framing that keeps beginners out of trouble: ExPASy tools answer narrow questions precisely. What is the molecular weight, where will trypsin cut, is there a kinase domain, is there a template for a model. They do not answer broad questions such as "what does this protein do in the cell" in one submission. You assemble that answer yourself, tool by tool, and then you test it. ## Frequently Asked Questions ### What are ExPASy bioinformatics tools? They are the web-based protein analysis tools and databases served by ExPASy, the Expert Protein Analysis System operated by the SIB Swiss Institute of Bioinformatics, covering tasks such as physicochemical property calculation, motif scanning, similarity search, proteolytic cleavage prediction, and homology modeling [1]. ### Do I need to install anything to use ExPASy? No. ExPASy runs entirely in a web browser, and every tool in this guide works from a pasted sequence or accession number, which is a large part of why it has served as the entry point for bioinformatics since 1993 [1]. ### Which ExPASy tool should I run first on a new protein? Run a similarity search first to establish identity, then ProtParam for basic physicochemical properties, then ScanProsite for domains and motifs, and then SWISS-MODEL if structural insight would help, following the order used in published protein characterizations [10], [6]. ### Why do my ProtParam molecular weights disagree with my gel? ProtParam computes the mass of the unmodified, unprocessed polypeptide chain, so signal peptides, propeptides, glycosylation, phosphorylation, proteolytic cleavage, and expression tags all shift the observed size, and you should compare against the exact construct and processing state you are actually working with. ### Can ExPASy translate a DNA sequence into a protein? Yes. The Translate tool converts a nucleotide sequence into amino acid sequences in all six reading frames, and you pick the open reading frame that starts with methionine and lacks internal stop codons. ### How much should I trust SWISS-MODEL models? Trust them in proportion to their template quality and sequence coverage, read the server's quality estimates before using the coordinates, and remember that published workflows treat model evaluation with separate tools as a required step after modeling [6]. ### Do I need programming to learn bioinformatics with ExPASy? No. All core ExPASy tools are point-and-click web applications, which is why undergraduate bioinformatics courses have used them alongside BLAST and KEGG to teach complete, hypothesis-driven analyses without any coding [3]. ### Does ExPASy cover anything besides protein sequences? Yes. Beyond the protein core, the site hosts glycoinformatics resources through the Glycomics@ExPASy tab, which centralizes databases and tools connecting glycan data with protein-focused resources [4], and if your work moves into [mass spectrometry-based proteomics](/knowledge/bioinformatics/mass-spectrometry-based-proteomics-data-analysis-pipelines-and-tools) data analysis, free tools such as FragPipe and DIA-NN are a sensible next step beyond sequence-level prediction [11]. ## Sources 1. [ExPASy: The proteomics server for in-depth protein knowledge and analysis.](https://pubmed.ncbi.nlm.nih.gov/12824418/) 2. [The Swiss-Prot protein knowledgebase and ExPASy: providing the plant community with high quality proteomic data and tools.](https://pubmed.ncbi.nlm.nih.gov/15707838/) 3. [An inquiry into protein structure and genetic disease: introducing undergraduates to bioinformatics in a large introductory course.](https://pubmed.ncbi.nlm.nih.gov/16220142/) 4. [Glycomics@ExPASy: Bridging the Gap.](https://pubmed.ncbi.nlm.nih.gov/30097532/) 5. [Bioinformatics Resources for the Study of Glycan-Mediated Protein Interactions.](https://pubmed.ncbi.nlm.nih.gov/35129172/) 6. [A bioinformatics approach to characterize a hypothetical protein Q6S8D9_SARS of SARS-CoV.](https://pubmed.ncbi.nlm.nih.gov/37037461/) 7. [[Cloning of TmHQTs gene from Taraxacum mongolicum and bioinformatics analysis and functional validation of HQTs proteins].](https://pubmed.ncbi.nlm.nih.gov/42392734/) 8. [[Bioinformatics analysis and prokaryotic expression of Strongyloides stercoralis serine protease inhibitor 1].](https://pubmed.ncbi.nlm.nih.gov/37455094/) 9. [[Properties comparing and evolutionary analysis of MEF2 of Homo sapiens based on bioinformatic methods].](https://pubmed.ncbi.nlm.nih.gov/21951798/) 10. [Identification and characterization of novel antimicrobial peptides from Camelus dromedarius: a combined bioinformatics and experimental study.](https://pubmed.ncbi.nlm.nih.gov/41659852/) 11. [Navigating the Mass Spectrometry-Based Proteomic Data Using Free Computational Tools.](https://pubmed.ncbi.nlm.nih.gov/40920679/)