April 24, 2018 by Dale Yuzuki
American Association for Cancer Research 2018 Lung & Colon Cancer Panel in a CRC Cohort Poster Interview with Dr. Jianfeng Xiao
At the recent American Association for Cancer Research (#AACR18), a Pillar Biosciences affiliate Shanghai Zhengu Biotech Limited’s lead scientist Jianfeng Xiao, Ph.D. presented a poster entitled “Evaluation of ONCO/Reveal Lung & Colon Cancer Panel in a CRC Cohort and Effect of DNA Damage in Variant Detection”.
Here’s a video interview where he discusses his results.
A transcription of this AACR 2018 poster interview follows:
Dale: Here I am at AACR 2018 in Chicago Illinois, with Dr. Jienfeng Xiao of Shanghai Zhengu Biosciences, which is Pillar Biosciences’ affiliate in mainland China. Jianfeng, thank you for taking the time to talk with me about your poster at AACR. What can you tell me about it?
Dr. Xiao: I am going to present research that we did using the ONCO/Reveal Lung and Colon Cancer Panel, to test more than 200 colorectal cancer (CRC) samples.
Dale: And this is the ONCO/Reveal Lung and Colon Cancer Panel from Pillar Biosciences, is that correct?
Dr. Xiao: Yes, and together with their PiVAT bioinformatics pipeline that Pillar Biosciences also provides.
Dale: And what did the 200 samples of data reveal?
Dr. Xiao: At the beginning, it was just to test the performance of the panel, so we tested more than 200 of these CRC samples and based on the data all samples were successfully run, and we summarized the most frequently mutated genes: TP53, KRAS, PIK3CA, FBXW7 and PTEN. And CNV was also observed in some genes, like EGFR, MET, ERBB2 and KRAS.
Dale: I understand from the poster you took a subset, is that right, from these samples?
Dr. Xiao: Yes, we looked at the distribution of variant frequencies, and we found an abnormally high number of variants with low allele frequencies, below 2%. And most of them, I mean more than 70% of those low-frequency variants, are C to T (cytosine to thymine) or G to A (guanine to adenosine).
Dale: And that was a clue, right? When you have C to T transversions… that’s DNA damage, correct?
Dr. Xiao: So we realized those low-frequency variants might be false-positives, caused by DNA damage. So we then did research on the DNA damage.
Dale: I see, so how many samples then did you pull out?
Dr. Xiao: We selected 27 samples, and those samples were treated with DNA repair enzymes before the library preparation, and the matched fresh-frozen samples for them were also tested.
Dale. I see, then what did you find?
Dr. Xiao: And we compared the variant list, and after DNA repair the number of low frequency variants in FFPE were lower, but there were still more compared to the fresh-frozen samples. So we concluded that the DNA repair could help, but not eliminate, the DNA damage.
Dale: So from FFPE samples, you were able to get to what low allele frequency with the DNA repair enzyme?
Dr. Xiao: The PiVAT pipeline can get an error LOD (limit of detection) down to 1%, it’s okay for fresh-frozen; but from FFPE samples, our recommendation based on our research is that it should be set to be no less than 2%.
Dale: I see, so you were able to get to 2% allele frequency from FFPE without any kind of barcoding, is that correct, with the PiVAT pipeline, can call that low?
Dr. Xiao: Yes, yes. I know that (Thermo Fisher’s) AmpliSeq™ has an LOD of 5%, but in our study we do find true variants between 2% and 5%, lots of hotspot variants, for example PIK3CA and KRAS. Those variants are verified by ARMS-PCR. (ARMS stands for Amplification Refractory Mutation System, commonly abbreviated PCR-ARMS.)
Dale: Thank you very much for sharing this with us today.
Here is more information on the ONCO/Reveal Lung & Colon Cancer Panel, and if you’d like to download the Pillar Bioscience affiliate Shanghai Zhengu AACR 2018 poster it is available.