1. Advances on the MIC's n6mA

2. Hemi-methylated AT sites in the MAC

PhD Project

Biggest question: How does the cell recognize the Scan-RNA independant IES ?

2 independant hypothesis:

• Some motifs or specific nucleotide composition

• Permanently epigenetic marking - DNA methylation

PacBio SMRT sequencing

- Inter-Pulse Duration (IPD) ⇈ if n6mA

- Compared with Control or in-sillico control --> IpdRatio

- Each molecule --> analyzed independantly

- Each nucleotide --> Analyzed independantly

- Each strand --> analyzed independantly (then paired)

> 99% accuracy

~15% error

Strategy:

1:200

Random sampling

PacBio SMRT

• 1 out of 200 comes from the MIC
• 1/6 of MIC inserts will carry an IES
• 1/2 of IESs are just wrongly excised
• 30% of the remnants are scanRNA dependant

Only a few remaining: ~ 10 to ~100 sequences

If 100% carry a methylation pattern, this is enough

Sorting

Deduced origin

MIC DNA

Alignment of consensus

MAC, MIC,

Mito...

Analysis

Report n6mA

5mC, 4mC

Re-alignment

Available data at day0

Wild type:

• Vegetative Cell (HTVEG)
• During autogamy  (after 2h - HT2)
• During autogamy  (after 6h - HT6)
  • Asynchronous cultures / Quite blurred state
     

Silencing of methylase candidates:

• Si/MAB (identified since then - probably a histone methylase)
   
• MT proteins:
  • Si/MT2
  • Si/MT1A-1B
  • Si/MT1A-1B-2
    
     
• NM proteins:
  • Si/NM4-9-10
  • Si/NM9-10
  • Si/NM4

Determining Se and Sp

• Sensitivity, Specificity of n6mA detection ?
  • No data for our approach in the litterature
    • Short inserts
    • Sequel vI.0
  • Can't afford a real benchmarking

• Paramecium is fed with E.coli !
  • ~ 100% n6mA in GATC
  • EcoK1 methylation well documented
  • Benchmark on E.coli ?

Found:

• Se ~ 93%  = P(D+ | M)
• Sp ~ 99.9% = P(D- | NM)

Detected level VS real level

$$p= [ Se \cdot \pi  + (1 - Sp) \ (1 - \pi) ] \cdot N$$

True positives

False positives

p : Number of D+

N: Number tested

$$\pi = \frac{\frac{p}{N} - 1 + Sp}{Se-1+Sp}$$

--> From imperfect detections, allows to estimate the fraction of nucleotides truly methylated

$$\pi$$ is the true proportion of n6mA

Quantitative estimation of the n6mA in the MAC

Mostly in AT sites (~90%)

"Lots" are symmetrical (~80%)

Quantitative estimation of the n6mA in the MIC

Quantitative estimation of n6mA in the MIC  (details)

Distribution of the detections

HTVEG

MT2

etc...

--> Some molecules carry all the detections, in sym-A*T

Very likely to be sequences comming from the MAC

What else could we explore ?

In the MIC:

• Other type of DNA methylation
• Kinetic signal around the IES
• That's it

What else could we explore ?

In the MAC: Hemimethylation

Thanks !

Output example

DNA n6mA

Analysis of Paramecium tetraurelia

Enfin !

Per experiment comparaison (1)

Per-experiment comparaison (2)

In the vegetative MAC

• ~95% of the methylation locates in AT dinucleotides in the MAC(*)

  • slightly lower in the MIC (5 to 5 points less)

  • True in any condition
     

• 75% of the methylation in an AT dinucleotide is actually symetrically modified, independantly from being in the MAC or the MIC(*)

Kept in MIC and MAC (All conditions)

(*) Linear equation / idQv20

Outside AT sites

Kept in the MAC for all experimental conditions

Impossible to tell in the MIC (not enough sequences)

In the MAC

Per genome comparaison

mDNA

42% GC

rDNA

38% GC

How could IES be recognized ?

• Weak consensus TAYAG
  • Not sufficient to recognize the IES
  • Degenerated TC1-Mariner TE ?
• Periodic distribution of size

~ 100% TA bounded

Small-RNA (~30% of IESs)

Output example