1/n Thanks to NAR Cancer @NAR_Cancer_EIC, for publishing our recent finding along with Dr. Sudhir Nair @ACTREC_TMC We have a long-standing interest in studying pathogens associated with human cancer.@India_Alliance @DBTIndia @the_hindu @RPrasad12

11 of n: We emphasize the need for stringent bioinformatics approaches to ensure accurate cancer microbiome profiling from liquid biopsy. Thanks to ICMR (@ICMRDELHI), DHR (@DeptHealthRes) for the funding. And,@DAEIndia for the transformative agreement with.@SpringerNature for OA.

10 of n: Offnote, efforts from research groups have shown plasma microbiome signatures as a promising approach in detecting colorectal, pancreatic (metagenome-based) & lung cancer (WGS-based). Our study, along with others, however, suggests several more nuances to be overcome.

9 of n: It highlights key limitations in our study: low microbial DNA abundance, risk of overestimation by standard tools, need for better microbial enrichment, and cleaner databases. So we now know what not to do — and that’s a crucial lead!.

8 of n: We thus concluded that under our experimental set up neither CGP nor WES — in their current form — can reliably detect a microbial DNA signature in plasma to identify lung cancer. But this study is still relevant.

7 of n: IPD2 revealed that many of those microbial reads detected from WES were likely false positives — originating from misalignment, contaminated databases, or misclassified human reads. Thus, the organ-specific microbial signature didn't hold up. Another negative result!.

6 of n: But as always, scientific interpretation is based on rigor in analysis, we compared two computational tools with statistical rigor: the widely used Kraken 2 and our in-house developed IPD2 — optimized for stringency and accuracy to detect microbial reads from blood.

5 of n: WES, on the other hand, picked up much higher microbial reads — and some even clustered in a way that looked like lung-specific microbial signatures. We were excited, as this suggested a breakthrough to identify tumors of unknown origin from a liquid biopsy.

4 of n: Microbial reads in these studies are usually off-target in an effort to target human genes. We find that CGP provided very few microbial reads — not adequate to find an organ-specific microbial signature. Here's the first negative result!.

First, we used CGP (~1K genes) across 261 plasma samples in collab with Drs. Bharde, Khandare, & Shafi from Then we went a step further: WES on 40 samples — a technical feat given the sparse amount of cfDNA with Mr. Goswami & Dr. Veldore from @4baseCare.

2 of n: But this “negative” result adds significant insights to cancer diagnostics! Vichitra, Supriya, @desaisanket12, & team set to ask: can circulating microbial DNA in the blood help identify the site of the primary tumor? We explored this using two genomic approaches.

Pleased to share our study led by Dr. K Prabhash, @VanitaNoronha from @TataMemorial, and me from @UnivofDelhi in collab with @1cellAi & @4baseCare where we found… nothing! Published in NPJ System Biology & Applicant (@Nature_NPJ).

9 of n: It highlights key limitations in our study: low microbial DNA abundance, risk of overestimation by standard tools, need for better microbial enrichment, and cleaner databases. So we now know what not to do — and that’s a crucial lead!.

8 of n: We thus concluded that under our experimental set up neither CGP nor WES — in their current form — can reliably detect a microbial DNA signature in plasma to identify lung cancer. But this study is still relevant.

7 of n: IPD2 revealed that many of those microbial reads detected from WES were likely false positives — originating from misalignment, contaminated databases, or misclassified human reads. Thus, the organ-specific microbial signature didn't hold up. Another negative result!.

6 of n: But as always, scientific interpretation is based on rigor in analysis, we compared two computational tools with statistical rigor: the widely used Kraken 2 and our in-house developed IPD2 — optimized for stringency and accuracy to detect microbial reads from blood.

5 of n: WES, on the other hand, picked up much higher microbial reads — and some even clustered in a way that looked like lung-specific microbial signatures. We were excited, as this suggested a breakthrough to identify tumors of unknown origin from a liquid biopsy.

4 of n: Microbial reads in these studies are usually off-target in an effort to target human genes. We find that CGP provided very few microbial reads — not adequate to find an organ-specific microbial signature. Here's the first negative result!.

First, we used CGP on ~1K genes across 261 plasma samples in collab with Drs. Bharde, Khandare, & Shafi from Then we went a step further: WES on 40 samples — a technical feat given the sparse amount of cfDNA with Mr. Goswami & Dr. Vedore from @4baseCare .

But this “negative” result adds significant insights to cancer diagnostics! Vichitra & Supriya from KP and my lab set to ask: can microbial DNA in the blood (liquid biopsy) help identify the site of the primary tumor? We explored this in lung cancer using two genomic approaches.

Stepping into the mind of a 10-year-old! Audvik’s take on our family picnic—pure, unfiltered…. ⁦@DuttShilpee⁩