HRD 1. a simple and reliable HRD detection method

A simple and reliable HRD The test method of

Hungry for knowledge No, BUG

background

Shengxin Xiaozao

Ovarian cancer is the gynecological malignant tumor with the highest mortality ,70% Of patients with ovarian cancer were in advanced clinical stage . In recent years , Poly adenosine diphosphate ribose polymerase （poly ADP ribose polymerase,PARP） The advent of inhibitors has brought great changes to the treatment of ovarian cancer , A series of high-level evidence-based medical evidence shows that complete remission is achieved after initial treatment or platinum sensitive recurrence treatment （complete response,CR） And partial remission （partial response,PR） Post application PARP Inhibitors can significantly prolong progression free survival in patients with ovarian cancer （progression free-survival,PFS） Time , Maintenance therapy has become a new model of ovarian cancer treatment [1].

BRCA1/2 Genes, tumor suppressor genes , stay DNA Damage repair 、 Normal cell growth and other aspects play an important role . Their mutations can inhibit DNA Normal repair ability after damage , Cause homologous recombination defects （homologous recombination deficiency, HRD）, namely BRCA Loss of function or other homologous recombination related bases Due to mutation or loss of function , Breaking the double chain DNA Repair Cannot be repaired by homologous recombination （homologous recombinant repair,HRR）, Eventually lead to cancer .

Biomarkers

HRD It can be judged by homologous recombination related gene mutation detection and gene scar detection . at present PARP In clinical trials of inhibitors HRD The latter method is often used , It can not only detect BRCA state , It can also analyze the type of genetic instability .[1]

How many people will benefit ？

HRD Detection can make PARP The inhibitor sensitive population accounts for 20% Left and right BRCA The mutant population expanded to account for 50% Left and right HRD Positive people .[1]

How to detect ？

Myriad Genetics myChoice HRD The detection is mainly aimed at genomic instability 3 The number of indicators in tumor samples was comprehensively scored , namely LOH、 Telomere allele imbalance （telomeric allelic imbalance,TAI）、 Large segment migration （large-scale transition,LST）. If score ≥42 Minute or BRCA1/2 mutation , Is defined as HRD positive ; If score ＜42 branch , And BRCA It's wild type , Is defined as HRD negative .

This method is also internationally tested HRD One of the more mature methods . Compared to using HRD dependent panel, Calculate the score through mutation , It is easier to standardize , Although it cannot reflect the epigenetic changes , There is a certain false negative . Because it is easy to implement , It can be combined with some means of genome-wide exon detection . If your patient already has external data , Then let's have a look .

Tools

scarHRD It's a R package , It can be used NGS（WGS、WES） The data of , To assess the LOH、LST And TAI. by the way , What it goes is Myriad Genetics myChoice Platform approach . Next , Let's introduce how to install and use it . Address ：（https://github.com/sztup/scarHRD）

01

install

Excellent network users

library(devtools)
install_github('sztup/scarHRD',build_vignettes = TRUE)

If your network is excellent , Just install it directly ……

Users with poor network

library(devtools) 
install("/path/scarHRD-master/")

If your network or luck is the same as mine , Then I can only think about the rut , from github Download source code , hold zip Unpack the package to the server , Follow the example above , Compile your unzipped directory .

Tips

The above one is a kind of R Offline installation mode of package , Mother doesn't worry about me anymore 404 了 .

02

Input

Input file

scarHRD The input file of is another R package Sequenza The intermediate output of “*small.seqz.gz” file , If you have this R Bag's experience , Then everything will be very easy .

If not , First run and compare , hold sort and dedup Of bam The file came out , At the same time Sequenza Description of .

If you can't run even if you are right , You can pay attention to my official account. , Some resources can also be pushed to you in the future .

03

function

command

#R Run in

library("scarHRD")
scar_score("/pathto/test1.small.seqz.gz",reference = "grch38", seqz=TRUE)

Tips

1）reference It can be “grch38” perhaps “grch37”

2） If your alignment reference sequence is “1”、“2”,“3” This format , Remember to deal with it small.seqz.gz file , Put the chromosome ID Change to “chr1”、“chr2”、“chr3”, This format , Otherwise, an error will be reported .

The result shows that

## Determining HRD-LOH, LST, TAI

##      HRD Telomeric AI LST HRD-sum
## [1,]  25           35  33      93

stay R See such a paragraph in , congratulations , succeed .

Just like the guide ,42 Points and above are positive HRD, The first three numbers are the scores of three sub items . Unweighted sum , The last number is the total score .

04

Software evaluation

I'm making a fool of myself

From the relevant references , This method can better identify HRD The data of .TCGA In the data ,95% Of HRD Tumor patients scored at 42 above .HRD Score of tumor patients , The lower quartile is 55. The overall performance is excellent .

I also ran some data myself , The expectation is HRD Positive patients are much higher than the threshold , The scores of several control patients are not too high , The difference is quite good …… Naked eye big data series

reference ：

[1] oophoroma PARP Guidelines for clinical application of inhibitors

[2] Patterns of genomic loss of heterozygositypredict homologous recombination repair defects in epithelial ovarian cancer.

All the contents of this article , Including text 、 picture 、 Audio 、 video 、 Software 、 Program 、 And web page layout design are collected online . Welcome to forward , Please register the source for reprint .

Welcome to scan the code and pay attention to Shengxin Xiaozao , Exchange and learn the original student letter content ~