Explain spatial transcriptomics

6.5.11 · Biology › Systems Biology & Frontiers

Overview

Spatial transcriptomics ek revolutionary technique hai jo gene expression (kaun se genes RNA mein transcribe ho rahe hain) measure karti hai aur saath mein tissue ke andar cells ki spatial location bhi preserve karti hai. Traditional RNA sequencing (RNA-seq) tissue ko cells ki ek "soup" mein grind kar deti hai, jisse saari spatial information destroy ho jaati hai. Spatial transcriptomics is sawaal ka jawaab deti hai: "Tissue mein KAHAN pe kaun se genes active hain?"

Yeh kyun important hai:

• Cells ka behaviour unke neighbours aur microenvironment par depend karta hai
• Cancer jaisi diseases spatial patterns create karti hain (tumor core vs. invasive edge)
• Development spatial gradients follow karta hai (embryos mein head vs. tail)
• Same cell type alag-alag tissue locations mein different gene expression rakh sakta hai

Spatial context ke bina, hum critical biology miss kar dete hain: cells kaise organize hote hain, communicate karte hain, aur functional tissue architecture kaise banate hain.

Core Concepts

The Central Problem

Traditional RNA-seq workflow:

1. Tissue ko dissociate karo → single cell suspension
2. Har cell se RNA extract karo
3. Transcripts ko sequence aur count karo
4. ❌ Spatial information hamesha ke liye lost ho jaati hai

The innovation: RNA molecules ko ek spatially-barcoded surface par capture karo, jahan har location par ek unique DNA barcode hota hai. Jab tum RNA+barcode ko saath sequence karte ho, tum har transcript ko uski tissue location tak trace kar sakte ho.

Yeh Kaise Kaam Karta Hai: First-Principles Derivation

Method 1: Spatial Barcoding (Visium/10x Genomics)

Step 1: Ek spatially-barcoded slide banao

• Ek glass slide par millions of capture probes print kiye jaate hain
• Har probe ki structure hoti hai: [Spatial Barcode] - [UMI] - [polyT tail]
• Probes spots mein organize hote hain (55 μm diameter, ~10 cells per spot)
• Har spot ka ek unique spatial barcode hota hai

PolyT kyun? mRNA molecules mein polyA tails hoti hain. PolyT unhe specifically bind karta hai.

Step 2: Fresh-frozen tissue ko slide par rakho

• Tissue section (10 μm thick) directly barcoded spots par rakha jaata hai
• Tissue ko permeabilize kiya jaata hai → mRNA molecules diffuse ho kar nikaLti hain aur nearest capture probes se bind karti hain

Step 3: Reverse transcription spatial information capture karta hai

mRNA-AAAA (from tissue)
     ||||
Probe-TT-[UMI]-[Spatial Barcode]-[on slide]
          ↓ Reverse transcription
        cDNA-[UMI]-[Spatial Barcode]

Yeh step critical kyun hai: Spatial barcode ab mRNA ki cDNA copy se physically attached hai. Jab bhi hum cDNA ko slide se sequencing ke liye nikaalte hain, woh apni location information saath le jaata hai.

Step 4: Sequence aur decode karo

• Slide se saara cDNA nikalo, amplify karo, sequence karo
• Har read mein hota hai: Gene identity + UMI + Spatial barcode
• Reads ko spot coordinates par map karo → XY positions ke saath gene expression matrix

Spot $i$ jo position $(x_i,y_i)$ par hai, usme gene $g$ ka expression level hai:

$E_{g,i}={\sum }_{j\in \text{UMIs at spot }i}1({\text{UMI}}_j\text{ maps to gene }g)$

UMIs count kyun karte hain, reads kyun nahi? PCR amplification ke dauran, ek mRNA molecule kaafi saari identical copies banata hai (PCR duplicates). Unique Molecular Identifier (UMI) ek random barcode hai jo amplification se pehle add kiya jaata hai. Duplicates ka same UMI hota hai, isliye hum true molecule count paane ke liye unique UMIs count karte hain.

Resolution: Har spot = ~5-10 cells, isliye yeh near-cellular resolution hai, single-cell nahi.

Method 2: In Situ Sequencing

RNA ko slide par capture karne ki jagah, ise tissue ke andar hi sequence karo.

Step 1: Structure preserve karne ke liye tissue fix karo

• Formaldehyde se proteins/RNA ko crosslink karo
• RNA molecules apni original positions mein lock ho jaati hain

Step 2: Tissue ke andar mRNA → cDNA reverse transcribe karo

• Padlock probes use karo jo target RNA sequences ke around circularize ho jaate hain
• Rolling circle amplification (RCA) create karo → bright fluorescent spots

Step 3: Tissue mein synthesis se sequence karo

• Ek-ek base par fluorescent nucleotides add karo
• Har addition ke baad tissue ki image lo
• Har XY position par colour record karo → sequence decode karo

Step 4: Sequences ko genome se align karo

• Har fluorescent spot = ek specific XY coordinate par ek mRNA molecule
• Subcellular resolution (~100 nm) achieve hoti hai lekin limited gene panel ke liye

Trade-off: In situ methods mein spatial resolution zyada hoti hai lekin yeh sirf 100-1000 pre-selected genes measure kar sakti hain (whole transcriptome nahi).

Worked Examples

Scenario: Mouse hippocampus (memory formation region) mein spatial gene expression map karo.

Approach: 10x Visium use karo (55 μm spots, whole transcriptome).

Results interpretation:

• Spot A (CA1 pyramidal layer mein): Camk2a, Grin1 ka high expression (excitatory neuron markers)
• Spot B (dentate gyrus mein): Prox1, Calb2 ka high expression (dentate granule cells)
• Spot C (white matter mein): Mbp, Plp1 ka high expression (oligodendrocyte myelin genes)

Yeh step kyun matter karta hai: Yeh expression patterns known anatomy confirm karte hain. Alzheimer's study karne wala researcher ab pooch sakta hai: "Kaun se spatial regions abnormal gene expression dikhate hain?"

Calculation: Agar spot A mein Camk2a ke liye 500 UMIs hain aur spot mein ~8 cells hain, toh average expression per cell ≈ 500/8 = 62.5 UMI/cell. (Yeh approximate hai kyunki spot ke saare cells neurons nahi hain.)

Scenario: Breast cancer biopsy mein ek tumor mass normal tissue se ghira hua dikhta hai.

Approach: Tumor center → edge → normal tissue tak gene gradients measure karne ke liye spatial transcriptomics.

Finding:

• Tumor core: MKI67 (proliferation), VEGFA (angiogenesis) ka high expression
• Invasive edge: MP9 (matrix degradation), SNAI1 (epithelial-mesenchymal transition) ka high expression
• Immune infiltrate zone: CD8A, PDCD1 (T cells, exhaustion markers) ka high expression

Spatial kyun matters hai: Bulk RNA-seq in teeno zones ko average kar deta, jo tumor invasion ki critical biology ko chhupaata. Spatial data reveal karta hai ki edge par invasive cells ka expression program core mein proliferating cells se alag hai.

Therapeutic insight: Agar tum proliferation (tumor core) ko target karo lekin invasion genes (tumor edge) ko ignore karo, toh cancer invasive cells se recur ho sakta hai jo maare nahi gaye.

Common Mistakes & Misconceptions

Kyun sahi lagta hai: "Spatial" word precise lagtaa hai.

Reality: Zyaadatar commercial platforms (Visium) mein spots hote hain jisme 5-10 cells hoti hain. Tum un cells ka average expression measure karte ho.

Fix:

• Visium ≈ 55 μm spots = near-cellular resolution
• In situ methods (MERFISH, seqFISH) = subcellular resolution (~0.1-1 μm) lekin fewer genes
• True single-cell spatial = newer methods jaise Xenium, CosMx, ya computationally Visium spots deconvolving

Steel-man: Confusion isliye hoti hai kyunki single-cell RNA-seq exist karta hai, aur "spatial" ek upgrade lagtaa hai. Lekin spatial barcoding trade-off karta hai: tum location gain karte ho, lekin single-cell resolution lose karte ho (jab tak tum specialized in situ methods use nahi karte).

Kyun sahi lagta hai: Agar spot A mein gene X ke liye 1000 UMIs hain aur spot B mein 500 UMIs, toh spot A mein 2× zyaada expression hai, right?

Reality: Total UMI counts spot ke hisaab se vary karte hain kyunki:

• Tissue thickness variation
• Cell density differences (10 cells vs. 5 cells per spot)
• RNA quality (degraded regions mein fewer counts hote hain)

Fix: Compare karne se pehle Normalize karo: ${\tilde{E}}_{g,i}=\frac{E_{g,i}}{{\sum }_{g^{\prime }}{E}_{g^{\prime },i}}\times 10^4$

Yeh counts per 10,000 mein convert karta hai (RNA-seq mein TPM ki tarah), jisse spots comparable ban jaate hain.

Yeh step essential kyun hai: Normalization ke bina, tum soch sakte ho ki spot A gene zyaada express karta hai, jabki actually usne sirf zyaada total RNA capture kiya.

Kyun sahi lagta hai: Hum gene expression ko living cells mein ek dynamic process ki tarah sochte hain.

Reality: Tissue ko fresh-frozen (Visium) ya formaldehyde-fixed (in situ methods) hona zaroori hai. Dono cells ko kill karte hain. Hum preservation ke waqt RNA ka ek snapshot measure kar rahe hain.

Fix: Spatial transcriptomics steady-state RNA levels capture karta hai jo tissue harvest hone ke waqt exist karte the. Dynamic processes ke liye (jaise circadian rhythms), tumhe tissue ko multiple time points par harvest karna hoga.

Implication: Tum spatial patterns aur function ke beech correlation dekhte ho, causation nahi. Agar gene X region Y mein high hai, toh yeh prove nahi hota ki X, Y ka phenotype cause karta hai—yeh cell type ka marker ho sakta hai, environment ka response ho sakta hai, etc.

Active Recall Questions

#flashcards/biology

Spatial transcriptomics kya hai? :: Ek technique jo gene expression (RNA levels) measure karti hai aur saath mein tissue ke andar cells ki XY coordinates preserve karti hai, gene activity ka ek spatial map create karti hai.

Memory Aids

Socho: GPS navigation tumhe batata hai "kaun si stores MAP par KAHAN hain." Spatial transcriptomics tumhe batati hai "tissue mein KAHAN kaun se genes active hain."

Socho tumhare paas ek badi Lego city hai (yeh tumhara tissue hai, jaise brain ya skin ka tukda). Har Lego brick ek cell hai, aur har cell ke paas ek recipe book (genes) hai jo batata hai ki kya banana hai.

Ab, agar tum saare Lego bricks ko ek blender mein daalo aur poochho "kaun si recipes use ho rahi hain?", toh important information lose ho jaayegi: kya red bricks (fire trucks) fire station ke paas hain? Kya blue bricks (police cars) police station ke paas hain?

Spatial transcriptomics ek photo lene jaisi hai Lego city ki upar se aur har brick ko label karna ki woh abhi kaun si recipe padh raha hai. Ab tum patterns dekh sakte ho: "Oh! Saari fire truck recipes fire station ke paas active hain!" ya "Uh oh, villain's lair mein weird recipes hain jo wahan honi nahi chahiye."

Real biology mein, isse hume diseases samajhne mein madad milti hai. Cancer cells (villain's lair) center mein alag recipes rakh sakti hain vs. edge par jahan woh invade kar rahi hain. Agar hum sirf blended soup dekhte, toh hum miss kar dete ki edge cells "walls todne" wali recipes use kar rahi hain jabki center cells "jaldi multiply hone" wali recipes. Cancer cure karne ke liye, hume DONO rokna hoga, aur spatial transcriptomics humein dikhati hai ki har problem KAHAN ho rahi hai.

Connections & Further Learning

Prerequisites:

• 6.1-DNAstructure-and-replication
• 6.2.3-Transcription-and-gene-expression
• 6.3.5-RNA-sequencing-basics

Related Concepts:

• 6.5.8-Single-cell-RNA-seq
• 6.5.9-Immunofluorescence-microscopy
• 6.5.10-FISH-fluorescence-in-situ-hybridization

Advanced Topics:

• 6.5.12-Spatial-proteomics
• 6.5.13-Computational-deconvolution-of-spatial-data
• 6.5.14-Spatial-multi-omics

Clinical Applications:

• 7.2.5-Tumor-microenvironment-mapping
• 7.3.2-Brain-region-specific-gene-expression

Summary

Spatial transcriptomics gene expression profiling mein revolution laati hai "kaun se genes KAHAN active hain?" ka jawaab deke. Yeh spatially-barcoded capture (Visium: whole transcriptome, near-cellular resolution) ya in situ sequencing (MERFISH/seqFISH: selected genes, subcellular resolution) use karta hai. Key insight: same identity wale cells alag-alag gene programs rakh sakte hain apni tissue location aur neighbours ke hisaab se.

Critical points:

1. Spatial barcodes RNA ko location coordinates se physically tag karte hain
2. UMIs true molecules count karte hain, PCR duplicates nahi
3. Spots compare karne se pehle normalization zaroori hai
4. Resolution vs. gene coverage ek trade-off hai
5. Spatial gradients, zones, aur microenvironments discover karna possible hota hai jo bulk RNA-seq ke liye invisible hain

Yeh technique cancer biology, neuroscience, developmental biology, aur immunology ko transform kar rahi hai, yeh reveal karke ki gene expression tissue architecture kaise create karta hai.

Concept Map