Tag: Artificial Intelligence

Free course · 11 chapters · 16 lectures · Urdu/Hindi AlphaFold Tutorial for Beginners: Proteins and Protein Structure Prediction, Chapter by Chapter Start with what a protein actually is. Finish by predicting a 3D structure on AlphaFold Server and working out what it does. Every chapter has my recorded lectures inside it, so you can read and watch side by side. By Dr. Muhammad Aammar TufailCodanicsUpdated 3 November 2026About 30 min read Subscribe on YouTube What is AlphaFold?AlphaFold is an artificial intelligence system from Google DeepMind that predicts the 3D structure of a protein from its amino acid sequence. It reached near-experimental accuracy in 2020, its creators shared the 2024 Nobel Prize in Chemistry, and its free database now covers more than 200 million proteins. This course teaches you to use it from zero, in 11 chapters and 16 video lectures. AlphaFold at a glance DeveloperGoogle DeepMind (AlphaFold 3 with Isomorphic Labs) BreakthroughAlphaFold 2 at CASP14, 2020; published in Nature, 2021 Latest versionAlphaFold 3, May 2024: proteins with DNA, RNA, ligands and ions RecognitionNobel Prize in Chemistry 2024 (Demis Hassabis and John Jumper, shared with David Baker) Input → outputAmino acid sequence → 3D coordinates (PDB/mmCIF) plus confidence scores (pLDDT, PAE) Free accessAlphaFold Database (200M+ structures) and AlphaFold Server, for non-commercial use 11chapters, read in order 16video lectures (4h 52m) 3hands-on assignments 200M+structures in the AlphaFold Database In 2024 the Nobel Prize in Chemistry went partly to Demis Hassabis and John Jumper of Google DeepMind for AlphaFold. For fifty years, working out the 3D shape of a single protein could take a PhD student years in the lab. AlphaFold does a good first attempt in minutes. Most AlphaFold tutorials skip straight to the tool and assume you already know structural biology. My students kept asking for something that starts from zero, in Urdu, and ends with analysis they could put in a thesis. So I recorded a full lecture series and wrote this guide to go with it. Treat it like a short book: one chapter at a time, with the lectures embedded where they belong. Chapters What is a protein?What are the four levels of protein structure?What is AlphaFold, and how do you use the AlphaFold Database?What is the difference between AlphaFold 2 and AlphaFold 3?How do you read a protein structure?How do you predict a protein structure with AlphaFold Server?How does AlphaFold work? The papers and the masterclassHow do you find a protein’s function? GO, KEGG and COGHow do you predict protein function from structure?Case study: what does AlphaFold show about p53?What can’t AlphaFold do? Glossary of terms All 16 lectures in one list FAQ How to study this: Chapters 1 to 4 take about an hour and need nothing but a browser. From Chapter 5 onward, keep a structure viewer open next to the lecture. Chapter 6 needs a Google account for AlphaFold Server. Do the three assignments when you reach them; they are where the learning sticks. Advertisement Chapter 1What is a protein?Lecture 01 · about 10 min of videoQuick answerA protein is a chain of amino acids that folds into a specific 3D shape. There are 20 standard amino acids, joined by peptide bonds, and their order (the sequence) is set by a gene. The shape a protein folds into decides its job, such as speeding up a reaction, carrying oxygen or fighting infection.In this chapterWhat proteins are made of, and why there are exactly 20 building blocksHow a gene turns into a protein sequenceHow to read a protein sequence and a FASTA fileWhere to find any protein’s sequence online Lecture 01 · 9:48Lecture 01: Introduction to Proteins & Protein Sequences. What proteins are, the 20 amino acids, and how to read a protein sequence and FASTA file. Your body contains tens of thousands of different proteins, and almost everything that happens in a cell is done by one of them. Enzymes speed up chemical reactions. Haemoglobin carries oxygen. Antibodies recognise viruses. Insulin tells cells to take up sugar. Most medicines work by sticking to a protein and changing what it does. The building blocks: amino acids A protein is a chain of amino acids. Every amino acid has the same backbone (an amino group, a central carbon and a carboxyl group) and a different side chain. Twenty standard side chains are used by every living thing. Their chemistry decides where each one ends up once the protein folds: GroupAmino acids (one-letter code)Where they like to be Hydrophobic (water-hating)G, A, V, L, I, M, F, W, PBuried in the core, away from water Polar, unchargedS, T, C, N, Q, YOn the surface, or in hydrogen bonds Positively chargedK, R, HOn the surface; often touching DNA, which is negative Negatively chargedD, EOn the surface; often in active sites Amino acids link head to tail through peptide bonds. A chain can be a few dozen amino acids long or several thousand. Each link in the chain is called a residue, a word you'll see in every AlphaFold output. From gene to protein The order of amino acids is written in DNA. A gene is copied into messenger RNA, and the ribosome reads that RNA three letters (one codon) at a time, adding one amino acid per codon. This is the central dogma of molecular biology, and it sets up the whole course: DNAthe gene→mRNAa working copy→Protein chainamino acids in order→3D shapethe fold→Functionthe job it does AlphaFold works on the middle arrow: from protein chain to 3D shape. Chapters 8 to 10 deal with the last arrow, from shape to function. Reading a protein sequence Bioinformatics writes sequences with the one-letter codes, so the start of the human haemoglobin alpha chain is MVLSPADKTNVKAAW…. Sequences are stored in FASTA files: one header line starting with >, then the letters. Almost every tool in this course, AlphaFold included, takes FASTA as input. >sp|P69905|HBA_HUMAN Hemoglobin subunit alpha MVLSPADKTNVKAAWGKVGAHAGEYGAEALERMFLSFPTTKTYFPHFDLSHGSAQVKGHGKK VADALTNAVAHVDDMPNALSALSDLHAHKLRVDPVNFKLLSHCLLVTLAAHLPAEFTPAVHA SLDKFLASVSTVLTSKYR Every known protein sequence lives in UniProt, with an accession ID such as P69905 for the protein above. That ID is how you'll look a protein up in the AlphaFold Database later. Key takeawaysA protein is a chain of 20 kinds of amino acids, joined by peptide bonds.Side-chain chemistry (hydrophobic, polar, charged) decides where each residue sits in the folded protein.Gene → RNA → protein sequence → 3D shape → function. AlphaFold predicts the shape.Sequences are stored as FASTA and found in UniProt by accession ID. Next: Chapter 2 · What are the four levels of protein structure? → Chapter 2What are the four levels of protein structure?Lecture 02 · about 13 min of videoQuick answerProtein structure has four levels. Primary structure is the amino acid sequence. Secondary structure is local shapes such as α-helices and β-sheets. Tertiary structure is the full 3D fold of one chain. Quaternary structure is how several chains fit together, as in haemoglobin’s four chains.In this chapterThe four levels of protein structureWhat makes a protein fold, and what happens when it misfoldsHow structures are measured in the labWhy predicting a structure was a 50-year problem Lecture 02 · 12:48Lecture 02: Types & Levels of Protein Structure. The primary, secondary, tertiary and quaternary levels of protein structure, and why folding matters. A newly made chain doesn't stay a loose string. In milliseconds to seconds it folds into a specific 3D shape, and that shape is what lets it do its job. Biologists describe the shape in four levels. AlphaFold's output only makes sense once these are clear. LevelWhat it isHeld together byExample PrimaryThe amino acid sequencePeptide bondsMVLSPADKTN… SecondaryLocal shapes: α-helices and β-sheetsHydrogen bonds in the backboneHelices in keratin (hair) TertiaryThe full 3D fold of one chainHydrophobic core, side-chain contacts, disulfide bondsMyoglobin QuaternarySeveral chains packed togetherThe same forces, between chainsHaemoglobin (4 chains) Why proteins fold The main driving force is the hydrophobic effect. Water-hating side chains pack together in the middle, like oil drops in water, and the charged and polar ones stay on the surface. Hydrogen bonds then lock helices and sheets in place. Some proteins get help from "chaperone" proteins that stop them clumping while they fold. When folding goes wrong, the results can be serious. Misfolded proteins that stick together are linked to Alzheimer's and Parkinson's disease, and one wrong amino acid in haemoglobin causes sickle cell disease. Shape is not a detail; it is the protein. Measuring a structure in the lab MethodHow it worksCatch X-ray crystallographyGrow a crystal of the protein and shoot X-rays at itMany proteins refuse to crystallise NMRRead atom positions from magnetic signals in solutionWorks best for small proteins Cryo-EMFreeze the protein and image thousands of copies with an electron microscopeExpensive equipment; best for large complexes All experimental structures go into the Protein Data Bank (PDB), which holds over 200,000 of them. That sounds like a lot until you compare it with the hundreds of millions of known sequences. Most proteins had no structure at all. Could we just predict the shape from the sequence? In 1969 Cyrus Levinthal pointed out that a protein can't find its fold by trying every possibility, because the number of possible shapes is astronomically large, yet proteins fold reliably within seconds. So the information must be in the sequence. Reading it out became known as the protein folding problem, and it stayed unsolved for half a century. Key takeawaysPrimary = sequence, secondary = helices and sheets, tertiary = one folded chain, quaternary = several chains.The hydrophobic effect drives folding; misfolding causes disease.X-ray, NMR and cryo-EM measure structures, but slowly and expensively.Sequences far outnumber known structures. That gap is what AlphaFold filled. Next: Chapter 3 · What is AlphaFold, and how do you use the AlphaFold Database? → Advertisement Chapter 3What is AlphaFold, and how do you use the AlphaFold Database?Lecture 03 · Lecture 04 · about 23 min of videoQuick answerAlphaFold is an AI system from Google DeepMind that predicts a protein’s 3D structure from its amino acid sequence. The free AlphaFold Protein Structure Database, run with EMBL-EBI, holds more than 200 million predicted structures. Search it by protein name or UniProt ID, then check the pLDDT confidence colours before you trust any region.In this chapterHow AlphaFold cracked the folding problemHow to search, view and download a structure from the AlphaFold DatabaseHow to read pLDDT and PAE confidence scoresAssignment 1 Lecture 03 · 22:12Lecture 03: AlphaFold Database & Tool for Protein Structure Prediction. How to search, view and download predicted 3D structures from the AlphaFold Protein Structure Database, and read pLDDT. Since 1994, structure prediction groups have competed every two years in CASP, a blind test where you predict structures that have been solved but not yet published. Progress was slow for decades. At CASP14 in 2020, AlphaFold 2 scored a median above 90 out of 100 on the main accuracy scale, close to experimental quality for most targets. The organisers said the problem was, in large part, solved. DeepMind and EMBL-EBI then ran AlphaFold on nearly every known protein and published the results free as the AlphaFold Protein Structure Database. It now holds more than 200 million predicted structures. In the lecture I use it live. The steps are: Go to alphafold.ebi.ac.uk and search by protein name, gene name or UniProt accession. Pick the right organism. "Insulin" alone returns dozens of species. Look at the 3D model. Its colours show confidence (explained below). Check the PAE plot under the model. Download the PDB or mmCIF file to open in a viewer. How much should you trust it? pLDDT Each residue gets a pLDDT score from 0 to 100: AlphaFold's own estimate of how right it is about that part of the model. The database colours the structure by this score: Very high pLDDT > 90Trust the backbone and most side chains.Confident 70 – 90Backbone is usually right; side chains less so.Low 50 – 70Treat with caution. Often flexible loops.Very low < 50Do not interpret. Often disordered regions. PAE (predicted aligned error) answers a different question: if I line the model up on residue A, how far off might residue B be? Low PAE between two domains means AlphaFold is sure how they sit relative to each other. High PAE means each domain may be fine on its own while the angle between them is a guess. On the plot, dark green blocks along the diagonal usually mark separate, well-defined domains. A common beginner mistake is to describe a long orange tail as a "loop" in a thesis. Very low pLDDT usually means the region is disordered, so it has no single fixed shape in the cell. Say that instead. Assignment 1: test your database skills Lecture 04 · 0:48Lecture 04: Assignment Alert! Test Your Knowledge. Assignment 1: search proteins in the AlphaFold Database, interpret pLDDT and describe their folding. Search the proteins given in the lecture on the AlphaFold Database.Record the pLDDT for each, and say which regions you trust and which you don’t.Download the structure files (PDB) and describe the secondary and tertiary folding you see.Submit your answers in the format given in the lecture. Spend the most time on step 2. Being able to say "this part I trust, this part I don't" is what separates people who use AlphaFold from people who screenshot it. Key takeawaysAlphaFold 2 reached near-experimental accuracy at CASP14 in 2020.The AlphaFold Database gives free structures for 200M+ proteins, searchable by UniProt ID.pLDDT is per-residue confidence: above 90 is very high, below 50 is very low and often disordered.PAE tells you whether the arrangement of domains can be trusted. Next: Chapter 4 · What is the difference between AlphaFold 2 and AlphaFold 3? → Chapter 4What is the difference between AlphaFold 2 and AlphaFold 3?Lecture 05 · about 6 min of videoQuick answerAlphaFold 2 (2021) predicts the structure of proteins. AlphaFold 3 (2024) predicts complexes of proteins with DNA, RNA, small-molecule ligands and ions, using a new Pairformer and diffusion-based architecture. Use AlphaFold 3 through AlphaFold Server when your question is about binding.In this chapterWhat AlphaFold 3 can do that AlphaFold 2 can’tHow the two models differ insideWhich one to use for your project Lecture 05 · 5:44Lecture 05: AlphaFold 2 vs. AlphaFold 3 (In-Depth Comparison). AlphaFold 2 vs AlphaFold 3 compared: architecture, DNA, RNA and ligand prediction, and use in drug discovery. AlphaFold 2 was about single proteins. But real biology happens when a protein meets something: a strand of DNA, a drug molecule, a metal ion, another protein. AlphaFold 3, published in Nature in May 2024 by Google DeepMind and Isomorphic Labs, models those complexes in one go. AlphaFold 2AlphaFold 3 Published2021, Nature (Jumper et al.)2024, Nature (Abramson et al.) PredictsProteins; complexes via AlphaFold-MultimerProteins, DNA, RNA, ligands, ions and some modified residues together InsideEvoformer and structure module, leaning heavily on the multiple sequence alignment (MSA)Pairformer and a diffusion module that builds atom positions directly, with less reliance on the MSA How you use itOpen-source code, AlphaFold DB, ColabFoldFree AlphaFold Server (non-commercial); code for academic use, weights on request Best for"What does this protein look like?""How does this protein bind its drug, DNA or partner?" The multiple sequence alignment is the trick behind both. AlphaFold gathers related sequences from other species. If two positions always change together across evolution, they are probably touching in 3D. AlphaFold 2 relied on this signal heavily. AlphaFold 3 still uses it, but the diffusion module, similar to the method behind AI image generators, refines a noisy cloud of atoms into a structure. For drug discovery this is a big deal: you can ask where a small molecule sits in a pocket, not just what the empty protein looks like. Even so, AlphaFold 2 is still the right tool for many jobs. The database is built on it, and ColabFold runs it quickly for free. Key takeawaysAF2 predicts proteins; AF3 predicts proteins together with DNA, RNA, ligands and ions.Both use evolutionary signal from related sequences (the MSA).Use AF3 (AlphaFold Server) when the question is about binding; AF2 is fine for a single protein. Next: Chapter 5 · How do you read a protein structure? → Chapter 5How do you read a protein structure?Lecture 06 · Lecture 07 · about 22 min of videoQuick answerRead a protein structure one element at a time: find the α-helices (spiral ribbons), β-sheets (flat arrows), loops (thin tubes) and domains (compact units). Use a free viewer such as Mol*, ChimeraX or PyMOL, and colour AlphaFold models by B-factor to see their confidence.In this chapterHow to make sense of a ribbon diagramSpotting helices, sheets, loops and domainsWhich free viewer to useAssignment 2 Lecture 06 · 19:08Lecture 06: Reading Protein Structures & Its Elements. How to read a 3D protein structure: alpha-helices, beta-sheets, loops and domains in PyMOL, ChimeraX or Mol*. The first time you open a 3D protein, it looks like a plate of jalebi: ribbons twisting everywhere. The trick is to stop staring at the whole thing and pick out four elements one at a time. α-helices, drawn as spiral ribbons. A rigid coil with about 3.6 residues per turn. β-strands, drawn as flat arrows. Strands lying side by side form a β-sheet; the arrow points along the chain. Loops and turns, the thin tubes connecting them. Binding often happens here, and pLDDT often drops here too. Domains, compact units that fold on their own. Big proteins are often several domains on a string, each with its own job. Free viewers ViewerGood for Mol*No install, runs in the browser. The same viewer the AlphaFold DB and PDB use. UCSF ChimeraXFree for academics; has built-in AlphaFold tools and pretty figures. PyMOLThe classic for publication figures and scripting. In any viewer, first colour by secondary structure to see helices and sheets. Then colour by B-factor: AlphaFold stores pLDDT in the B-factor column, so this brings back the confidence colours and shows you which parts to ignore. Assignment 2: analyse a sample structure Lecture 07 · 3:19Lecture 07: Read Protein Structures (Sample & Assignment). A live walkthrough of a sample protein structure, followed by Assignment 2 on structural analysis. In this lecture I break down a real model live, then hand it over to you: Open the structure of the protein given in the lecture in any viewer.Find and count its helices, sheets and loops.Identify its domains (UniProt or InterPro will help you name them).Submit a short structural analysis with screenshots. Key takeawaysRead a structure element by element: helices, sheets, loops, domains.Mol* needs no install; ChimeraX and PyMOL are best for figures.Colour by B-factor to see AlphaFold confidence in any viewer. Next: Chapter 6 · How do you predict a protein structure with AlphaFold Server? → Get PhD, postdoc and bioinformatics updatesFollow the Codanics channel on WhatsApp for funded positions and new lectures in AI, data science and bioinformatics.Join on WhatsApp Chapter 6How do you predict a protein structure with AlphaFold Server?Lecture 08 · Lecture 09 · about 40 min of videoQuick answerSign in to alphafoldserver.com with a Google account, paste your amino acid sequence as a protein entity, add any DNA, RNA, ligand or ion partners, and submit. In minutes you get five ranked models with pLDDT, PAE, pTM and ipTM scores, which you can download as mmCIF files.In this chapterWhen to use AlphaFold Server instead of the databaseRunning a prediction step by stepReading pTM, ipTM and the five modelsA small protein to practise on Lecture 08 · 16:39Lecture 08: AlphaFold Server Complete Guide & Live Example. AlphaFold Server step by step: submit your own sequence, run a prediction and read the results. The database covers proteins already in UniProt. What if your protein is a new mutant, a designed sequence, or bound to DNA or a drug? Then you run the prediction yourself on AlphaFold Server, which runs AlphaFold 3 in your browser. Sign in with a Google account. The server is free for non-commercial use, with a daily job limit. Paste your amino acid sequence as a protein entity. Add more entities (protein, DNA, RNA, ligand, ion) for a complex. Name the job and submit. Small jobs usually finish in a few minutes. Open the result: five ranked models, a pLDDT-coloured view, the PAE plot, and pTM and ipTM scores. Download the zip. It contains mmCIF files for any viewer. pTM scores the overall fold. ipTM scores the contacts between chains. As a rough guide, ipTM above 0.8 is a confident complex, and below 0.6 the predicted binding is probably wrong. Try it yourself: predict ubiquitinUbiquitin is small (76 amino acids), stable and very well studied, so it makes a good first job. Paste this as a protein entity:MQIFVKTLTGKTITLEVEPSDTIENVKAKIQDKEGIPPDQQRLIFAGKQLEDGRTLSDYNIQKESTLHLVLRLRGGYou should get a model that is almost all dark blue. Then open the experimental structure 1UBQ in the PDB and compare the two. Seeing them match is the moment most students start to trust the tool. The full workflow, A to Z Lecture 09 · 23:07Lecture 09: AlphaFold Server Tutorial (Protein Structure A to Z). A complete AlphaFold Server tutorial from an unknown sequence to a final 3D model, with advanced settings. This longer lecture does it the way you would for a thesis. I start with a sequence I know nothing about, find it, set up the job properly, use the advanced options (multiple copies of a chain, ligands, seeds), and interpret what comes back. AvoidRunning one job, screenshotting model 0 and calling it "the structure". DoCompare all five models. Where they agree, the model is likely right. Report pLDDT, PAE and ipTM with every figure. Key takeawaysUse the database for known proteins and the Server for new sequences or complexes.Each job gives five models; check how well they agree.pTM judges the fold, ipTM judges the interfaces. Below 0.6 ipTM, be sceptical. Next: Chapter 7 · How does AlphaFold work? The papers and the masterclass → Advertisement Chapter 7How does AlphaFold work? The papers and the masterclassLecture 10 · Lecture 11 · about 95 min of videoQuick answerAlphaFold learns from known structures and from evolution: related sequences in a multiple sequence alignment reveal which residues touch in 3D. AlphaFold 2 turns this into a structure with its Evoformer and structure module; AlphaFold 3 uses a Pairformer and a diffusion module. The original Nature papers explain both.In this chapterWhere to learn AlphaFold from its creatorsWhich papers to read, and how to read themA 93-minute masterclass bringing it all togetherAssignment 3 Running a tool doesn't make you a researcher. Knowing how it works, and where it breaks, does. This chapter sends you to the original sources and then pulls everything together in the longest lecture of the series. Assignment 3: the EMBL-EBI course and the Nature papers Lecture 10 · 1:21Lecture 10: Task Assignment! EMBL Course & AlphaFold Nature Paper. Assignment 3: the EMBL-EBI AlphaFold course and the AlphaFold Nature papers, including pLDDT and PAE. Complete the free EMBL-EBI AlphaFold online course.Read the AlphaFold 3 paper: Abramson et al., "Accurate structure prediction of biomolecular interactions with AlphaFold 3", Nature 2024.Write down how the method works, where it fails, and what pLDDT and PAE mean in the authors’ own words.Submit your summary as explained in the lecture. For background, also skim the original paper: Jumper et al., "Highly accurate protein structure prediction with AlphaFold" (Nature, 2021). That is the work the 2024 Nobel Prize recognised. On a first read, don't fight the methods section. Read the abstract, the figures and the limitations, then go back. The complete AlphaFold masterclass Lecture 11 · 1:33:38Lecture 11: AlphaFold Complete Guide: Understand Protein Structures A-Z. A 93-minute AlphaFold masterclass: AlphaFold 3 architecture and interpreting complex structures from A to Z. This 93-minute session joins Chapters 3 to 6 and goes deeper into AlphaFold 3: how the architecture works, how to set up complex jobs, and how to interpret a difficult structure from start to finish. Watch it in two sittings with a viewer open, and repeat each step on your own protein. Key takeawaysThe EMBL-EBI course is free and written by the people who run the AlphaFold DB.Read the AF2 (2021) and AF3 (2024) Nature papers: figures and limitations first.The masterclass is the best single revision of the whole AlphaFold part of the course. Next: Chapter 8 · How do you find a protein’s function? GO, KEGG and COG → Chapter 8How do you find a protein’s function? GO, KEGG and COGLecture 12 · about 37 min of videoQuick answerUse three databases. Gene Ontology (GO) describes a protein’s molecular function, biological process and location in the cell. KEGG maps it onto pathways and disease networks. COG places it in an evolutionary family. UniProt lists all three for known proteins.In this chapterWhy a structure alone doesn’t tell you a protein’s jobGene Ontology, KEGG and COG, and what each tells youHow to get annotations for known and new proteins Lecture 12 · 37:18Lecture 12: Understand Protein Function: GO, KEGG & COG Databases. How to find protein function with Gene Ontology (GO), KEGG pathways and COG orthologous groups. You have a sequence and a structure. Your supervisor's next question will be: "So what does it do?" Three databases answer it from different angles. DatabaseQuestion it answersWhat you get Gene Ontology (GO)What does it do, in which process, and where in the cell?Terms in three groups: molecular function, biological process, cellular component KEGGWhich pathway or disease network is it part of?Pathway maps (metabolism, signalling, disease) with your protein highlighted COGWhich evolutionary family does it belong to?Clusters of orthologous groups and a one-letter functional category Think of it this way: GO describes the job, KEGG shows the team the protein works in, and COG tells you its family tree. For known proteins, the UniProt entry already lists all three. For new sequences from your own genome or transcriptome, tools like eggNOG-mapper assign them in bulk. Key takeawaysGO = function, process, location. KEGG = pathways. COG = evolutionary family.UniProt collects all three for known proteins.For new sequences, use annotation tools such as eggNOG-mapper. Next: Chapter 9 · How do you predict protein function from structure? → Chapter 9How do you predict protein function from structure?Lecture 13 · Lecture 14 · about 31 min of videoQuick answerCompare your predicted structure with known structures using Foldseek or DALI, look for conserved pockets and active sites, and match domains in InterPro. Shape is conserved longer than sequence, so structure search often finds relatives that sequence search misses.In this chapterWhy shape is conserved longer than sequenceFinding pockets, active sites and similar structuresA reusable sequence → structure → function pipeline Lecture 13 · 27:32Lecture 13: Predict Protein Function from 3D Structures. Predicting protein function from 3D structure using active sites, binding pockets, Foldseek and DALI. Evolution changes sequences faster than shapes. Two proteins can share less than 20% of their sequence and still have the same fold and a similar job. So when a sequence search finds nothing, a structure search often still can. Structural search: upload your model to Foldseek (fast, searches the AlphaFold DB and PDB) or the DALI server. Hits with known functions give you a hypothesis. Pockets and active sites: look for clefts on the surface lined with conserved residues. Catalytic residues usually sit in high-pLDDT regions. Domains: match each domain to a family in InterPro. Everything here is a hypothesis to test, not a result. A structural match tells you what a protein could do. The whole pipeline in one picture Lecture 14 · 3:32Lecture 14: Sequence to Structure & Function Prediction (Easy Method). The full pipeline from sequence to AlphaFold structure to function prediction, in one short lesson. This short lecture joins the course into one workflow you can reuse on any protein: 1SequenceGet the FASTA from UniProt or your own data2StructureAlphaFold DB, AlphaFold Server or ColabFold3Check confidencepLDDT per residue, PAE between domains4CompareFoldseek or DALI against known structures5FunctionGO terms, KEGG pathways, COG groups Need to run hundreds of proteins without the server's daily limit? ColabFold runs AlphaFold 2 in Google Colab and is how many students get through large batches. Key takeawaysStructure is conserved longer than sequence, so structure search finds distant relatives.Foldseek and DALI compare your model with known structures.Sequence → structure → confidence → comparison → function annotation. Next: Chapter 10 · Case study: what does AlphaFold show about p53? → Chapter 10Case study: what does AlphaFold show about p53?Lecture 15 · about 14 min of videoQuick answerThe AlphaFold model of human p53 (UniProt P04637) shows a very-high-confidence DNA-binding domain and disordered, low-confidence tails. The common cancer mutations R175H, R248Q and R273H all sit in that DNA-binding domain, which explains why they disable the protein.In this chapterWhy p53 matters in cancerRunning the full workflow on a real proteinWhat cancer mutations look like on a 3D model Lecture 15 · 13:55Lecture 15: Understand Protein Function: Real-Life p53 Walkthrough. A real case study of the p53 tumour suppressor: AlphaFold structure, GO/KEGG function and cancer mutations. p53 is a transcription factor that stops a damaged cell from dividing, or tells it to die. Its gene, TP53, is mutated in roughly half of all human cancers, which is why p53 is called the guardian of the genome. In the lecture we take it through every step of the course: Domain map of human p53Schematic of the 393-residue p53 protein showing the disordered N-terminal transactivation and proline-rich regions, the high-confidence DNA-binding domain from residue 94 to 292 containing hotspot mutations R175, R248 and R273, the tetramerisation domain and the disordered C-terminus.Transactivation: residues 1–61TransactivationProline-rich: residues 64–92Proline-richDNA-binding domain: residues 94–292DNA-binding domainTetramer: residues 323–356TetramerC-term: residues 364–393C-termR175HR248QR273H194292393residueHuman p53 (393 residues), drawn to scale. Colours follow AlphaFold’s pLDDT scheme for each region of the database model: blue = high confidence, orange = very low (disordered). Red dots mark the three most common cancer hotspot mutations. Domain boundaries are approximate. StepWhat we find Sequence (Ch. 1)Human p53 is UniProt P04637, 393 amino acids. Structure (Ch. 3)In its AlphaFold model the central DNA-binding domain is dark blue (very high confidence). The long N- and C-terminal tails are orange. Those tails are disordered in real life, not badly predicted. Reading it (Ch. 5)The core is a β-sandwich, two packed β-sheets, which holds up the loops and a helix that contact DNA. Function (Ch. 8)GO terms point to DNA binding and to regulating apoptosis and the cell cycle. KEGG puts it at the centre of the p53 signalling pathway and many cancer pathways. MutationsThe common cancer hotspots, such as R175H, R248Q and R273H, are in the DNA-binding domain. On the 3D model you can see why: they either touch DNA directly or upset the fold that holds the contact surface in place. Key takeawaysp53 is mutated in about half of human cancers.Its AlphaFold model shows a confident core and disordered tails, a textbook example of reading pLDDT.Mapping mutations onto structure explains why some changes break the protein. Next: Chapter 11 · What can’t AlphaFold do? → Chapter 11What can’t AlphaFold do?Lecture 16 · about 2 min of videoQuick answerAlphaFold predicts one static shape, so it misses protein movement. It is poor at judging the effect of single mutations, cannot give a fixed shape for disordered regions, and can produce convincing but wrong complexes when ipTM is low. Important structures still need experimental support.In this chapterHow structure prediction is entering genome projectsThe limits you must mention in any report Lecture 16 · 2:07Lecture 16: Integrate Protein Structures in a Genome Assembly Pipeline?. How AlphaFold structure prediction can be integrated into genome assembly and structural annotation. The last lecture looks ahead. After a new genome is assembled and its genes predicted, annotation normally compares sequences against databases. For many genes in less-studied organisms, that finds nothing. Predicting structures for those proteins and searching them with Foldseek can assign functions that sequence search misses. This is called structural annotation, and it is becoming a normal step in genome projects. What AlphaFold can't do AlphaFold is a model, not an experiment. Keep these limits in mind before you write anything up: It predicts one static shape. Real proteins move, and many change shape when they bind something. Most single point mutations barely change the predicted structure, even when they destroy function. Don't use AlphaFold alone to judge whether a mutation is harmful. Disordered regions come out as low-confidence spaghetti. That is information, not failure. Complexes with low ipTM, or ligands placed in poorly predicted pockets, can look convincing and still be wrong. For a key structure in a publication, experimental evidence (or an existing structure in the PDB) is still expected. Key takeawaysStructure prediction now helps annotate whole genomes.AlphaFold gives one static snapshot; report confidence scores and limits with every figure. You've finished the course. What next? AlphaFold is one part of a bigger skill set. My full course, Bioinformatics ka Chilla, covers genetics, NGS data and analysis from basic to advanced. To turn your results into a paper, watch the free Scientific Writing playlist. For the bigger picture, see the free AI and data science roadmap. Glossary of AlphaFold and protein termsShort definitions of every technical term used in this courseAmino acidOne of 20 building blocks that link together to form a protein chain.ResidueA single amino acid within a protein chain. AlphaFold scores every residue.FASTAA plain-text format for sequences: a header line starting with “>”, then the letters.UniProtThe main database of protein sequences and annotations, with IDs such as P04637.PDBThe Protein Data Bank, the archive of experimentally determined 3D structures.CASPCritical Assessment of protein Structure Prediction, a blind test held every two years since 1994.MSAMultiple sequence alignment: related sequences lined up so evolutionary patterns can be read.pLDDTAlphaFold’s per-residue confidence score from 0 to 100. Above 90 is very high; below 50 is very low.PAEPredicted aligned error: AlphaFold’s expected position error between pairs of residues, used to judge domain arrangement.pTMPredicted TM-score: a 0–1 confidence for the overall fold of a model.ipTMInterface pTM: a 0–1 confidence for the contacts between chains in a complex. Above 0.8 is confident.DomainA compact part of a protein that folds on its own and often has its own function.Intrinsically disordered regionA part of a protein with no single fixed shape. Usually shows very low pLDDT.FoldseekA fast tool that searches for proteins with similar 3D structures.Gene Ontology (GO)A standard vocabulary for protein function, biological process and cellular location. All 16 lectures in one list The full playlist, plus every lecture grouped by chapter Full playlist: AlphaFold & Protein Structure Prediction Course (Urdu/Hindi). Plays all 16 lectures in order. #LectureLengthChapter 1 · What is a protein?01Introduction to Proteins & Protein Sequences9:48Chapter 2 · What are the four levels of protein structure?02Types & Levels of Protein Structure12:48Chapter 3 · What is AlphaFold, and how do you use the AlphaFold Database?03AlphaFold Database & Tool for Protein Structure Prediction22:1204Assignment Alert! Test Your Knowledge Assignment0:48Chapter 4 · What is the difference between AlphaFold 2 and AlphaFold 3?05AlphaFold 2 vs. AlphaFold 3 (In-Depth Comparison)5:44Chapter 5 · How do you read a protein structure?06Reading Protein Structures & Its Elements19:0807Read Protein Structures (Sample & Assignment) Assignment3:19Chapter 6 · How do you predict a protein structure with AlphaFold Server?08AlphaFold Server Complete Guide & Live Example16:3909AlphaFold Server Tutorial (Protein Structure A to Z)23:07Chapter 7 · How does AlphaFold work? The papers and the masterclass10Task Assignment! EMBL Course & AlphaFold Nature Paper Assignment1:2111AlphaFold Complete Guide: Understand Protein Structures A-Z1:33:38Chapter 8 · How do you find a protein’s function? GO, KEGG and COG12Understand Protein Function: GO, KEGG & COG Databases37:18Chapter 9 · How do you predict protein function from structure?13Predict Protein Function from 3D Structures27:3214Sequence to Structure & Function Prediction (Easy Method)3:32Chapter 10 · Case study: what does AlphaFold show about p53?15Understand Protein Function: Real-Life p53 Walkthrough13:55Chapter 11 · What can’t AlphaFold do?16Integrate Protein Structures in a Genome Assembly Pipeline?2:07 Codanics on YouTubeBioinformatics, data science, AI and research careers, in Urdu and Hindi.Subscribe Frequently asked questions What is AlphaFold in simple words?AlphaFold is an AI program from Google DeepMind that looks at a protein’s amino acid sequence and predicts the 3D shape it folds into. Scientists use those shapes to understand diseases and design drugs, without spending months determining each structure in the lab.Is AlphaFold free to use?Yes for academic and non-commercial work. The AlphaFold Protein Structure Database is free to search and download. AlphaFold Server, which runs AlphaFold 3, is free with a Google account but has a daily job limit. AlphaFold 3 code is available for academic use, with model weights on request from Google DeepMind.How accurate is AlphaFold?For many well-folded proteins, AlphaFold’s high-confidence regions (pLDDT above 90) are close to experimental accuracy. Accuracy drops for flexible loops, disordered regions, proteins with few relatives in sequence databases, and some complexes. Always check the pLDDT, PAE and ipTM scores.What is the difference between AlphaFold 2 and AlphaFold 3?AlphaFold 2 (2021) predicts protein structures, and protein complexes through AlphaFold-Multimer. AlphaFold 3 (2024) also predicts how proteins interact with DNA, RNA, small molecules and ions, using a Pairformer and a diffusion module instead of AlphaFold 2’s Evoformer-centred design.What does pLDDT mean in AlphaFold?pLDDT is a per-residue confidence score from 0 to 100. Above 90 is very high confidence, 70 to 90 is confident, 50 to 70 is low, and below 50 is very low. Very-low regions are often intrinsically disordered rather than wrongly predicted.Do I need a biology background to learn AlphaFold?No. Chapters 1 and 2 start from what a protein is and how it folds. Students from computer science and data science follow this course without a biology degree.Do I need coding skills or a GPU to use AlphaFold?No. The AlphaFold Database and AlphaFold Server both run in a web browser. Coding helps only if you want to run many predictions, for example with ColabFold in Google Colab.Can AlphaFold tell me what a protein does?Not directly. AlphaFold predicts shape. To infer function you compare the shape with known structures using Foldseek or DALI, look for pockets and active sites, and check annotations in GO, KEGG and COG. Chapters 8 to 10 cover this.Can AlphaFold predict the effect of a mutation?Not reliably. Most single amino acid changes barely alter AlphaFold’s predicted structure, even when they destroy the protein’s function. Use dedicated variant-effect tools and experimental data for that question.Is this AlphaFold course available in Urdu or Hindi?Yes. All 16 video lectures are in Urdu/Hindi with technical terms in English, which suits students in Pakistan and India. This written guide is in English so you can use it as notes while you watch. Don't miss the next courseFollow Codanics on WhatsApp for new lectures, PhD and postdoc openings, and AI and data science news.Join on WhatsApp Advertisement About the author. Dr. Muhammad Aammar Tufail is the founder of Codanics and completed his PhD in a Marie Skłodowska-Curie programme. He teaches bioinformatics, data science, AI and research skills to students in Urdu and Hindi. Sources: Jumper et al., Nature 596, 583–589 (2021); Abramson et al., Nature 630, 493–500 (2024); AlphaFold DB FAQ (pLDDT bands and colours); Nobel Prize in Chemistry 2024; EMBL-EBI AlphaFold course; UniProt entries P69905 and P04637; PDB entry 1UBQ. Lectures from the Codanics AlphaFold playlist.

Tag: Artificial Intelligence

AlphaFold Tutorial for Beginners: Free Course (16 Lectures)

Marie Curie PhD Positions 2026–27: All 141 Approved Doctoral Networks, Salaries in Rupees and How to Apply | Codanics

Study Abroad, PhD & Research: 16 Practical Lessons for Students | Codanics

LinkedIn Networking Guide for Pakistani AI & Data Science Students | Codanics

Web Scraping with Python | Codanics

A Guide to Prompt Engineering for Data Scientists & AI Developers in Pakistan | Codanics

What is MLOps? A Beginner’s Guide | Codanics

Top 10 benefits of using vscode as your code editor or IDE

Artificial Intelligence (AI): Aik Naye Zamane Ka Aghaz! 🌟

Data Science & AI: Judwa Bhaai Ya Masaledar Biryani ki Two Ingredients? 🍲

Quick Links